RSS
 

Posts Tagged ‘Face’

Sony’s new HVL-F28RM flash uses camera face detection for better portraits

18 Sep

Sony has just announced the HVL-F28RM flash. It’s designed to be compact and easy to use, but also comes with features that more advanced photographers will appreciate.

When combined with Sony’s a7C, a7S III, a7R IV and a9 II, the F28RM will use face detection information from the camera to better balance its output with the ambient lighting of the scene, as well as adjust white balance for a more natural look.

The flash angle can be adjusted up to 120 degrees and it’s powered by two AA-sized alkaline or NiMH batteries. Other features include radio wireless communication, a stronger metal hot shoe and dust and moisture resistance.

The HVL-F28RM will be available this winter and will retail for $ 249 USD ($ 329 CAD).

Press release:

Sony Electronics Introduces Alpha 7C Camera and Zoom Lens, the World’s Smallest and Lightest[i] Full-frame Camera System

New HVL-F28RM Compact Flash is also Announced

SAN DIEGO, CA – September 14, 2020 – Today, Sony Electronics Inc. announced several additions to an already impressive imaging lineup — the Alpha 7C full-frame camera (model ILCE-7C), the FE 28-60mm F4-5.6 (model SEL2860) zoom lens and HVL-F28RM flash.

The Alpha 7C is the world’s smallest and lightest[ii] full-frame body with uncompromising performance, featuring advanced AF (autofocus), high-resolution 4K video[iii] capabilities and more. When paired with the world’s smallest and lightest[iv] FE 28-60mm F4-5.6 standard zoom lens, this versatile combination delivers an experience unlike any other, maximizing portability and versatility without sacrificing any of the power of full-frame imaging. The HVL-F28RM flash allows users to broaden their photo expressions with outstanding compactness, and an intelligent light intensity control linked to camera face detection[v].

“We are committed to creating the best tools possible, based on the needs of our customers,” said Neal Manowitz, deputy president of Imaging Products and Solutions Americas, Sony Electronics. “The new Alpha 7C camera and FE 28-60mm F4-5.6 zoom lens pack many of our most advanced imaging technologies in a brand new design that is the smallest and lightest full-frame camera and lens system in the world. This opens up a new world of possibilities for creators, giving them the uncompromised power of a full-frame system in the palm of their hand.”

…

New HVL-F28RM: Compact Flash with Light Intensity Control Linked to Camera Face Detection[v]

The HVL-F28RM is a compact flash designed to match Sony’s mirrorless cameras for a compact, manageable system, and offers the type of reliable, stable performance that only a genuine Sony product can provide. When compared to the HVL-F32M, the HVL-F28RM features a 12 percent reduction in volume and 7 percent reduction in weight. This compact, easy-to-use flash unit delivers the capabilities and dependability to meet the needs of both professional and advanced amateur content creators.

The HVL-F28RM offers consistent GN28[xxv] light output, optimized light distribution and continuous flash performance that won’t interrupt the user’s workflow, as well as stable radio wireless communication and multi flash radio control. The new flash also features Sony’s newly introduced flash control linked to camera face detection[v] advanced technology. When used with a compatible camera, the balance between the light falling on the subject’s face and ambient light is evaluated to automatically adjust accurate white balance so that the subject’s face is rendered with natural, lifelike color. In addition, flash compensation, light ratio, and other detailed flash parameters can be controlled directly from a compatible camera[xxvi]. A camera custom key can be assigned to call up the flash parameter display so that adjustments can be made while looking through the viewfinder and gripping the camera. Flash parameters are shown in the selected camera display language.

A newly developed “Metal Shoe Foot with Rugged Side Frame”[xxvii] that also houses the unit’s electrical contacts offers improved resistance to physical shock and impact from all directions. The Multi Interface foot is fabricated from metal for higher strength. The HVL-F28RM also features a dust and moisture resistant[xxii] design. When the HVL-F28RM is mounted and locked onto the Alpha 7C, Alpha 7S III, Alpha 7R IV and Alpha 9 II, durability to dust and moisture is improved, even when used in challenging outdoor environments.

The HVL-F28RM also features simple, intuitive operation with minimal controls including +/- light level buttons, pairing button, test button and lock lever. Plenty of light is available for bounce applications. The flash angle can also be set as required via 0, 20, 40, 60, 80, and 120 degree click stops for easy positioning. The new flash also features a built-in wireless radio trigger for reliable flash triggering when mounted on a compatible camera[xxvi] and paired with an off-camera unit. When used as a transmitter, the HVL-F28RM can control up to 15 flash and/or receiver units in 5 groups[xxviii] at distances of up to 114 feet (35 meters)[xxix] for extraordinary lighting control and versatility. The HVL-F28RM is powered by two AA (LR6) alkaline or NiMH batteries. A fresh pair of alkaline batteries can provide power for up to 110 continuous flashes (1/1 manual flash with alkaline batteries)[xxix].

Pricing and Availability

The HVL-F28RM flash will be available this winter and will be sold for approximately $ 249.99 USD and $ 329.99 CAD. It will be sold at a variety of Sony’s authorized dealers throughout North America.

Exclusive stories and exciting new content shot with the new camera, lens and Sony’s other imaging products can be found at www.alphauniverse.com, a site created to educate and inspire all fans and customers of Sony ? – Alpha.

[i] An Alpha 7C with an FE 28-60mm F4-5.6 lens mounted. Among full-frame interchangeable-lens digital cameras, in combination with an interchangeable zoom lens. As of Sept. 2020. Sony survey.

[ii] Among full-frame interchangeable-lens digital cameras with optical in-body image stabilization mechanism, as of Sept. 2020. Sony survey.

[iii] A Class 10 or higher SDHC/SDXC card is required for XAVC S format movie recording. UHS speed class 3 or higher is required for 100 Mbps recording.

[iv] Among interchangeable zoom lenses for 35mm full-frame format digital camera bodies, as of Sept. 2020. Sony survey.

[v] This function is only compatible with Alpha 7C as of Sept. 2020

[xxii] Not guaranteed to be 100% dust and water resistant.

[xxv] 50 mm, at ISO 100 in meters

[xxvi] Visit Sony support webpage for functional compatibility information.

[xxvii] Design registration application pending.

[xxviii] In group flash mode. 3 groups (A-C) in TTL or manual flash mode.

[xxix] Sony internal test conditions.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Sony’s new HVL-F28RM flash uses camera face detection for better portraits

Posted in Uncategorized

 

Disney Research Studios demonstrates automatic face swapping with faster, cheaper AI

21 Jul

Disney Research Studios and ETH Zurich have published a study detailing a new algorithm that is able to swap faces from one subject to another in high-resolution photos and videos. Of note, this system is able to fully automate the face-swapping process, presenting the first instance of megapixel-resolution machine-generated imagery that is ‘temporally coherent’ and photo-realistic.

The new algorithm involves taking the face of a subject and modifying it using the face of another person, blending the two so that the face from one person is presented with the expressions and movements of another.

The system involves a multi-way comb network trained with images of multiple people, as well as a blending method that preserves contrast and lighting. ‘We also show that while progressive training enables generation of high-resolution images,’ the researchers say, ‘extending the architecture and training data beyond two people allows us to achieve higher fidelity in generated expressions.’

Key to the high level of quality is the ‘landmark stabilization algorithm,’ which Disney researchers describe as a ‘crucial’ aspect of dealing with high-resolution content. Though this isn’t the first instance of face-swapping in footage, the study points out that existing methods used to generate characters like the young Carrie Fisher in Rogue One are both time-intensive and quite expensive.

Artificial intelligence has the potential to change this, ultimately enabling creators to rapidly generate computer characters using live-action footage and input images of the target. Generating realistic faces remains a big problem, however, producing what is referred to as the ‘uncanny valley’ look that limits the use of this tech.

This makes Disney’s new technology particularly exciting, teasing a future in which creators will be able to generate photo-realistic, high-resolution, temporally-stable face swaps between two people. The researchers explain:

As our system is also capable of multi-way swaps — allowing any pair of performances and appearances in our data to be swapped — the possible benefits to visual effects are extensive, all at a fraction of the time and expense required using more traditional methods.

The study compares the face-swapping results from this new method to the results from existing algorithms, including DeepFaceLab and DeepFakes. Though the other algorithms were able to produce casually convincing results, they were unable to pass scrutiny and, in some cases, were either excessively blended or outright bizarre and uncanny.

This batch represents instances of failed face swapping

In comparison, the face swaps generated using the new method were realistic and maintained a high level of sharpness and detail at a 1024 x 1024 resolution, bypassing the soft, blurry results often seen when using DeepFakes. As well, the researchers note that DeepFakes has such heavy processing requirements that it was only able to generate a resolution of 128 x 128 pixels using an 11GB GPU.

When using morphable models, the researchers were able to increase the resolution to 500 x 500 pixels, but the results were typically unrealistic. Beyond that, the researchers were forced to train the conventional models for each pair of face swaps whereas the new algorithm could be simultaneously trained for all of the people used for the various face swaps.

However, the study points out that the new algorithm presents one big limitation also experienced by other, more conventional methods: the original head shape is maintained. Though the face swap may be very realistic, the face itself may not match the head shape properly, resulting in a generated character that looks a bit ‘off’ from what is expected.

Future research may result in a method for transferring the subject’s head shape in addition to their face, producing not only photo-realistic results, but also the correct overall appearance for a digitally-recreated actor. The biggest obvious use for this technology is in film and television, enabling studios to quickly and cheaply (relatively speaking) create 3D models of aging or deceased actors.

This technology joins a growing body of research on face-swapping and model-generating algorithms that focus on still images rather than videos. NVIDIA, for example, published a study in late 2018 that demonstrated the generation of photo-realistic portraits of AI models that involved source and target images of real people.

Around a year later, the same company published new research that performed a similar face swap, but one involving dogs instead of humans. We’ve already seen the use of these various AI technologies reach the consumer level — Let’s Enhance 2.0, for example, recently introduced a new feature that utilizes machine learning to reconstruct the faces of subjects in low-resolution images.

As for the new study from Disney Research Studios and ETH Zurich, the full paper (PDF) can be found on Disney’s website here.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Disney Research Studios demonstrates automatic face swapping with faster, cheaper AI

Posted in Uncategorized

 

Google Photos starts rolling out manual face tagging feature on mobile

03 Dec

As promised this past summer by Google Product Lead David Lieb, Google Photos has been updated to support manual face tagging. The feature first started rolling out to some users last week, according to Android Police, only days after XDA spotted signs of the new feature in an app teardown. The feature still has not arrived for all users, however.

Manual face tagging is a new Google Photos feature that builds upon the service’s existing face-detection algorithm. With this new tool, users are able to manually correct errors made by the algorithm and to also immediately tag images of new people who haven’t yet been identified by the app.

Users who have access to the feature note one big limitation with the new tool: the app does not allow users to tag faces that weren’t detected by the algorithm. As well, it isn’t yet clear whether manually tagging people and correcting mistakes will help improve the algorithm’s ability to detect those people in subsequent photos.

Users who have access to the new manual face tagging option can find the tool within the Google Photos app’s ‘Albums’ menu. Tap on ‘People & Pets,’ then tap on an image. Swipe up from the bottom of the screen to reveal the menu containing EXIF data, then swipe up again.

The person featured in the image will be listed under a section title ‘People.’ If you have access to the manual face tagging feature, you will see a new pen icon located next to the person within this ‘People’ section. Android Police notes that this feature is rolling out through a server-side update, meaning that users can’t manually update the app to get access to the new tagging option.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Google Photos starts rolling out manual face tagging feature on mobile

Posted in Uncategorized

 

Panasonic to sell remaining stake in semiconductor joint venture in face of ‘aggressive’ competition

03 Dec

Panasonic has announced it’ll be selling off its minority stake in its semiconductor joint venture for $ 250M to Taiwan’s Nuvoton Technology Corp after deciding it would need to invest more than it is prepared to do to compete and expand in the market.

Back in 2014, Panasonic offloaded a majority of its semiconductor unit to a joint venture with Isreali semiconductor manufacturer TowerJazz. The joint venture, which is owned 51% by TowerJazz and 49% by Panasonic, was initially believed to be sold as a whole, but TowerJazz has since confirmed in a statement that it will retain its majority stake and ultimate control of the operation following the transition:

‘TowerJazz, the global specialty foundry leader, clarifies following recent press releases in connection with the sale of Panasonic semiconductor business to Nuvoton that it will not sell its TPSCo shares and will maintain its 51% ownership and Board control in TPSCo.’

Part of the semiconductor business is involved with making imaging sensors for cameras and smartphones, as well as for numerous industrial purposes. It isn’t clear at the moment exactly how this will impact the company’s camera division or the upcoming 8K organic sensor planned for the 2020 Olympics, but all intellectual property and contracts are to be transferred to the buyer in June next year — a month before the start of the Olympics.

Panasonic says it’s tried to streamline its semiconductor business and that it has divested parts of the business already to make it less expensive to operate, but that it would need much more investment to expand the division and to compete in an aggressive market.

The sale may not have very much effect at all as the majority of Panasonic’s Lumix cameras use third-party sensors, and the majority of sensors made by the division being sold were for the automotive business and industrial applications. However, the division lists 16MP CMOS sensors for stills cameras and 20MP Super 35mm sensors for broadcast cameras in its offering. Whether the technology and manufacturing facilities for the 8K organic sensor are also part of the deal we have yet to discover.

We have contacted Panasonic for comment and to clarify what this might mean, if anything, for its camera business. We will update the article if we receive a statement.

Press release:

Announcement of the Transfer of the Semiconductor Business

OSAKA, Japan – Panasonic Corporation (hereinafter, the “Company”) announced that it will transfer (hereinafter, the “Transfer”) the semiconductor business mainly operated by Panasonic Semiconductor Solutions Co., Ltd. (hereinafter, “PSCS”), which is a 100% consolidated subsidiary company of Panasonic Equity Management Japan G.K.(hereinafter, “PEMJ”), a 100% consolidated subsidiary company of the Company, to Nuvoton Technology Corporation (hereinafter, “Nuvoton”), a Taiwan-based semiconductor company under the umbrella of Winbond Electronics Corporation group, and enter into the Stock and Asset Transfer Agreement (hereinafter, the “Agreement”) with this company. A decision was authorized by the Board resolution today.

1. Background and Purpose
The semiconductor business of the Company has shifted from the AV area to the automotive and industrial area over the last few years. The Company has positioned the “Sensing” technologies such as image sensors, and the “LiB Application” technologies such as IC for battery management and MOSFET for LiB battery circuits protection as the focus areas, and the Company has aimed to grow its business by consolidating resources in these areas.

In the meantime, in April 2014, the Company transferred the semiconductor wafer production process of the Hokuriku Plants (Uozu, Tonami, Arai) to the joint venture company formed with Tower Semiconductor Ltd., an Israel based foundry company. Furthermore, in June 2014, the Company transferred its semiconductor assembly plants in Singapore, Indonesia and Malaysia to UTAC Manufacturing Services Ltd. (hereinafter, “UTAC”) having its headquarter in Hong Kong. The Company has been strengthening its competitiveness by becoming an asset-light company, consolidating and eliminating its offices and production bases in both Japan and overseas for the mitigation of business risks.

However, the competitive environment surrounding the semiconductor business has become extremely severe due to aggressive expansion of competitors, huge investments in the focused area, and industry reorganization through M&A. In such an environment, the Company has come to believe that the even stronger business operation and the continuous investment is critical in order to achieve a sustained growth and expansion of the semiconductor business. Accordingly, it has concluded that the best option would be to transfer the business to Nuvoton, which highly appreciates the Company’s accumulated technical and product capabilities and therefore has a potential to lead stable growth by leveraging those capabilities.

2. About the Transfer
(1) Business restructuring before the Transfer: Just prior to the Transfer, the Company will restructure the semiconductor business as follows.

  • All shares of Panasonic Industrial Devices Systems and Technology Co., Ltd. (hereinafter, “PIDST”) and Panasonic Industrial Devices Engineering Co., Ltd. (hereinafter, “PIDE”), which are wholly-owned subsidiaries of PEMJ, will be handed over to PSCS by way of company split.
  • The semiconductor business-related intellectual property rights and certain business contracts held by the Company and/or the Company’s subsidiaries and the semiconductor business-related assets and debt of the Company will be handed over to PSCS by way of either company split or asset transfer.
  • All PSCS’s shares held by PEMJ will be handed over to a to-be-established, wholly-owned subsidiary of PEMJ (hereinafter, the “PSCS Holding Company”) by way of share transfer.
  • The semiconductor-related components (lead frame) business of PSCS will be handed over to a to-be-established, wholly-owned subsidiary of PEMJ by way of company split.

(2) Details of the Transfer: Upon completion of the business restructuring above, the Transfer will be carried out as per the details below with target effective date of June 1, 2020 (scheduled).

  • PEMJ will transfer all PSCS Holding Company’s shares to Nuvoton.
  • The business of Panasonic Industrial Devices Semiconductor Asia (an in-house company in charge of development and sales of semiconductors; hereinafter, “PIDSCA”) under Panasonic Asia Pacific Pte Ltd. (a Singaporean entity owned by the Company through its subsidiary; hereinafter, “PA”) will be handed over to Singapore- based entity owned by Nuvoton.
  • Certain facilities and inventories attributable to the semiconductor business of Panasonic Semiconductor (Suzhou) Co., Ltd. (hereinafter, “PSCSZ”) will be transferred to China-based entity owned by Nuvoton.

3. Other
The Agreement is based on the precondition of obtaining approvals from the authorities responsible for competition laws and other government agencies of the respective country and region. In addition, the planned date of the Transfer including business restructuring before the Transfer may differ significantly in light of the duration required for completing the procedures for obtaining approval and other procedures concerning permissions etc.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Panasonic to sell remaining stake in semiconductor joint venture in face of ‘aggressive’ competition

Posted in Uncategorized

 

NVIDIA’s latest AI project is ‘face swap’ for animals. Kind of…

30 Oct

NVIDIA researchers are back with another project that uses artificial intelligence to generate entirely new images from existing source images. Unlike past work that involved portraits of humans, however, this latest work — called GANimal — transforms an image of an animal into different animals, including other species.

Using an AI technique called generative adversarial networks (GANs), among other things, the researchers developed GANimal, an app that takes the expression of an animal from an image and recreates it on an image of a different animal. Examples include taking an expression from one breed of dog and replicating it on other dog breeds.

Though this is a fun example of the underlying technology, NVIDIA researchers say it could one day be put to use in more serious work. One given example of GANimal’s potential use is engaging filmmakers to shoot images of a tame animal doing stunts, such as a dog, and then using the AI to apply those movements onto a less tame animal, such as a tiger.

NVIDIA’s past artificial intelligence research includes an AI that can accurately scrub noise from images, generate portraits from source images, and transform simple sketches into photorealistic photos.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on NVIDIA’s latest AI project is ‘face swap’ for animals. Kind of…

Posted in Uncategorized

 

Facebook expands Face Recognition photo scanning, makes feature opt-in for new users

06 Sep

Facebook will no longer scan uploaded images for users’ faces by default, according to The Verge. The change will apply to new users who receive the Face Recognition setting as Facebook rolls it out globally over the next several weeks. The Face Recognition feature, which was first introduced in late 2017, will not be turned on unless the user chooses to enable it.

The facial recognition feature works by scanning images for users’ faces and alerting them about these images even if they’re not tagged in them. Users who receive one of these alerts can choose to tag themselves in the image, ignore it, or report the image when applicable.

In an update on the technology following the outcome of its federal appeal in August, Facebook has revealed that the facial recognition feature is rolling out to all users, but that they’ll need to manually enable it if they want the platform to scan other users’ images for their face. A notice in the user’s News Feed will alert that user when the feature becomes available on their account.

Users will be able to find the Face Recognition feature in their account’s Settings menu. Facebook users who currently have Face Recognition on their accounts can find instructions on disabling it here.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Facebook expands Face Recognition photo scanning, makes feature opt-in for new users

Posted in Uncategorized

 

The North Face flooded Wikipedia with product images to manipulate Google ranking

05 Jun

Last week, The North Face and ad agency Leo Burnett Tailor Made published a video detailing the company’s exploitation of Wikipedia as part of an ad campaign. The exploitation, according to the company’s video, involved swapping Wikipedia images of various destinations with new images that prominently featured The North Face gear.

The oddly boastful video puts forth the following question: How can a brand be the first on Google without paying anything for it? The rest of the video details how The North Face pulled off its stunt, including sending photographers to capture original images of people using the company’s gear in ‘adventurous’ locations.

The Wikipedia pages for these high-tourism destinations, the video notes, often appear at the top of the first Google search results page; the first images found on these pages are often at or near the top position on Google Image Search, as well.

As part of its manipulation, The North Face swapped the first images in these Wikipedia pages with its own original photos of the destinations — ones that prominently featured apparel, backpacks, and other products.

In pulling this stunt, the video brags that The North Face was able to manipulate Google Image Search into ranking its promotional content near the top of its results for these destinations.

A screenshot of the Guarita State Park Wikipedia page before North Face added its own photos.

The companies seemingly acknowledged the unacceptable nature of the activity to AdAge, reportedly stating the ‘biggest obstacle’ for the ad campaign was replacing the images ‘without attracting the attention of Wikipedia moderators.’ As well, the video at one point states that The North Face was ‘collaborating’ with Wikipedia in this effort, something Wikimedia Foundation addressed in a blog post.

A screenshot of the Guarita State Park Wikipedia page after North Face added its own photos.

The non-profit organization called the ad campaign an unethical manipulation of Wikipedia, saying, ‘They have risked your [the public’s] trust in our mission for a short-lived marketing stunt.’

‘Wikipedia and the Wikimedia Foundation did not collaborate on this stunt, as The North Face falsely claims,’ the non-profit states in its blog, comparing the ad campaign’s image manipulation to ‘defacing public property.’ As expected, Wikipedia proceeded to remove some of The North Face’s images from articles and to crop its logo out of other images.

Wikimedia Foundation said:

When The North Face exploits the trust you have in Wikipedia to sell you more clothes, you should be angry. Adding content that is solely for commercial promotion goes directly against the policies, purpose and mission of Wikipedia to provide neutral, fact-based knowledge to the world.

For its part, The North Face published a lackluster apology on Twitter, stating:


Articles: Digital Photography Review (dpreview.com)

 
Comments Off on The North Face flooded Wikipedia with product images to manipulate Google ranking

Posted in Uncategorized

 

Using Face Detection and Recognition in ACDSee Photo Studio Ultimate 2019

11 Dec

The post Using Face Detection and Recognition in ACDSee Photo Studio Ultimate 2019 appeared first on Digital Photography School. It was authored by Stacey Hill.

Are you a person who takes lots of photos of people? Perhaps you shoot weddings or events? Family portraits? Maybe you like to capture images of family and friends? Eventually, you end up with many images. Some are easy to sort through when they have only one person in them.

However, what about group photos? How do you tag/catalog/sort through those? Do you have to list out everyone’s name in the meta tags manually?

What if you don’t know all the names immediately? What if you find out later that Heather is actually Helen and you have to go back and change it?

Finally, how can you find all the images with a specific person quickly and easily?

Luckily, ACDSee Photo Studio Ultimate 2019 makes this task easy with the new Face Detection and Recognition capabilities.

How to set up Face Detection and Recognition

1.  Open the program and in ‘Manage’ mode, navigate to the desired folder where you have stored your images.

2.  Click on an image of the person you want to name and click on ‘View’ to open in View mode.

3.  Face Detection identifies the person by outlining their face.

4. If this outline box is not present, click on the ‘Show Face Outlines’ button (or Shift + B).

5. Once the face is selected, we need to apply the name. Click on the ‘Face’ Tool (or Shift + F) and a dark grey text bar pops up under the outline.

6. Click in the text bar and type the name of the person and press ‘Enter.’

7. If you select any other image with that person and open it in ‘View mode,’ it should automatically select their face and apply the name.

A key point regarding the naming structure you use is to put some thought into what naming convention you use and then keep it consistent. For example, something along the lines of: first name (space) last name or first name_last name. Doing so makes a difference later if you are using search parameters to find people.

What happens if you have more than one Joe Smith? What if you start off using only first names and then you have 3 x David, 4 x Michael, 2 x Louise all in the same wedding party? Being as specific as possible when naming addresses this issue.

How it becomes useful

Once the software has recognized the face and you have assigned a name to it, it detects that face amongst all your other photos. While it works across any images currently stored on your computer, the data is saved and applied to any new images you import onto your computer.

Therefore, you should only have to tag the person once. ACDSee remembers their name and applies it to any future image with their face in it.

Now you might want to search for all the images with Heather.

1. In ‘Manage’ mode, select ‘Catalog’ on the left-hand panel. The panel is broken up into different sections. At the top is ‘Categories.’

2. The second panel down is ‘People.’ All the names you have applied to your images are listed. Click on any name, and it goes through the database to pull out all the images with that face present in them.

3. In ‘Manage’ mode, you can ‘Quick Search’ by typing the name in the Quick Search box.

Searching for multiple people

1. Search for two or more people by holding down CTRL while selecting a second name in the People section of the Catalog panel. The software finds the images for those people. You can also utilize the ‘Easy Select’ arrows next to the names to select multiple people.

2. Find an image with two specific people in it together by typing both of their names in the ‘Quick Search’ bar as Person 1 + Person 2 to run a Boolean AND search. It is during this process that it’s essential to understand the naming convention you used originally.

In ‘Manage’ mode, select the folder you want to search in and click F3 to search (or right mouse click and select Search). Make sure you put ‘People’ in the ‘Categories’ section or it will search the entire database and potentially pull up other images.

3. You can also run a search from the ‘Catalog’ pane. In the ‘People’ section, select all the people you want to search for by using the ‘Easy-Select’ arrows or CTRL/SHIFT clicking. Click the gear icon on the People section header, and change the search type to ‘Match All.’

Managing name data

You can edit/change/remove the name data you have stored, which comes in handy if you have to update the spelling on one. Maybe you forgot you already had a ‘Sebastian’ in there, and you need to change one of them.

1. In ‘Manage’ mode, select ‘Tools’ from the top menu option.

2. From the drop-down menu, select ‘Manage People.’

3. A ‘People Manager’ box opens up with all the names you have saved. You can edit each one as needed by selecting them and using the bottom buttons.

Things to note:

1. The naming convention you use is important, so plan that out in advance.

2. If the face is not automatically detected, and you have to create it manually, the software will not further recognize it in Face Recognition. Also, note that if you use the Remove Faces or Redetect Faces command on an image, manual faces aren’t retained. The Rerun Face Detection option remembers them if you edit images.

3. Currently, there is no facility to import face recognition tags from other software (Picasa as an example). However, a search through the support forums has this listed on the ‘Potential Ideas for Future Updates’ list. It also appears to apply to the exporting of images from ACDSee as well, with the intention of retaining the face recognition tags.

4. There is no easy way to establish if there are currently any unnamed faces.

5. If the software has assigned the wrong name to someone, you can remove it with the Remove Faces function. This removes all face data from the selected image, not just the one wrong one.

6. To ensure a better success rate, you may need to manually select several images of one person so that the software can ‘learn’ that face with accuracy. You achieve better accuracy by naming as many faces in the (first) image as possible.

7. You can manually remove the names from incorrect selections and can rescan in ‘Manage’ mode via Tools | Redetect Faces. You need to correct the wrong name, rather than remove it, otherwise rescanning continues to return the wrong name over and over.

Conclusion

Face detection and recognition is a tool that can make life easier for a photographer with many images of people in their portfolio. The ability to assign a name to a person and have the computer run an algorithm to find all the other images is significantly faster than doing it manually.

To be able to search for images with a specific person (or range of people) becomes faster and more efficient as well.

Is it perfect?  If I am honest, not 100% all of the time. However, it is easy to use, easy to manage and does a pretty good job for most requirements. It could be useful for many other things that they may implement into the next version.

Right now, it is an effective time saver for the home photographer with photos of family and friends, through to commercial photographers with wedding/event shoots filling up the portfolio.

The previous 2018 generation of ACDSee was the first version that bought a range of features all together in one space. Thus, giving you the capability to manage and view files, edit Raw files, do creative editing with layers, all in one piece of software.

Leading with this Face Detection and Recognition, the 2019 iteration builds on that initial foundation by bringing specific functionality to boost capabilities even further. Thus, making for a compelling consideration for anyone looking to purchase editing software, especially when it is available via one-off perpetual license purchase.

 

 

 

The post Using Face Detection and Recognition in ACDSee Photo Studio Ultimate 2019 appeared first on Digital Photography School. It was authored by Stacey Hill.


Digital Photography School

 
Comments Off on Using Face Detection and Recognition in ACDSee Photo Studio Ultimate 2019

Posted in Photography

 

ACDSee Photo Studio Ultimate 2019 Review: Face Detection and Recognition

16 Oct

A question seen frequently on photography groups is “What software do I get to process my images in?”. There is the usual flurry of recommendations for the familiar choices and a few random ones thrown in. One option that doesn’t get mentioned as often as it should is ACDSee. In particular the Photo Studio Ultimate 2019 bundle is worthy of consideration for both beginners and more experienced users.

The 2019 version with the newly included Face Detection and Facial Recognition features is a step up from the previous 2018 program, indicating an intention towards AI-based digital asset management.

For anyone wanting a one-stop shop to manage, view, process RAW files, and edit with layers, etc, PLUS only having to pay once for a perpetual licence, ACDSee offers a compelling option in the marketplace.

My background is in Lightroom and Photoshop which is the basis for comparison in this review.

Let us assess this software from the point of view of what it offers a beginner.

CONTENTS

  1. Getting Started – installing and setting up
  2. Layout and Features
  3. Importing and Viewing Images
  4. Editing your RAW Image
  5. Advanced editing with layers
  6. New Features in 2019 version
  7. General Comments

1. Install and Setup

Setup and installation are fairly standard as per most software. ACDSee does require you to set up an account as part of the install process (it’s mandatory and cancels the install if you try to opt out), which then requires an extra registration step with an email confirmation. However, once sorted, no further registration is required. If you have registered before, you can use the previous login details.

It does allow you to choose which drive/directory/folder you want to install it into, as well as if you want to use a non-standard install path. As per the splash page below that opens on Startup—you can auto select the folder to open when the software starts.

Also new is the next screen, which helpfully shows you what the key functions and features are, and where to find them. Both of these can be turned off if desired. You can click on any of the words on the left panel and it will take you to the appropriate screen. Or click through on the NEXT button. Or close it.

 

Once you have navigated the splash pages, you will be taken to the Manage mode screen.

2. Layout and Features

ACDSee has five main modes in separate tabs for each function—Manage, Photos, View, Develop, and Edit.  There are some extra features but these are key ones used in general.

Summary of the features:

Manage mode has access to your computer, direction to find images where they are stored on the computer, and the default option is to view your images in thumbnail view (similar to Grid in LR). It shows EXIF data, histogram, and shot information for a particular image. You can colour code or rate images in Manage Mode.

Photos mode is similar to Manage. It allows a more comprehensive way of viewing image files on your hard drive, and you can drill down to specific day/month/year views.

View mode allows you to view a single image in full screen mode (similar to a single image view in LR) and has some basic editing functions included.

Develop mode is where you edit your RAW image files (similar layout and functions to LR or ACR).

Edit mode is where you can do advanced editing with layers (similar to PS).

There are also the 365 tab, Dashboard tab, and Messages tab. 365 is where you have access to your subscription information, if you opt for it. The Dashboard shows graphical data on image/camera information—if you want to know your most commonly used ISO setting, type, and number of files, it is visible here.

 

3.  Importing and Viewing Images

Importing is not required with ACDSee. The software will read folders directly off your computer, displaying and respecting its existing folder structure, just like Explorer. However, users can import off of external sources if they wish to achieve other organizational goals at the same time, such as culling, tagging, renaming, etc.

Once imported, you will then want to view them, cull, tag, and select the best ones for editing.

I have all my images stored on a NAS and it found those with no issues.  Above is the Manage page showing the hard drive directory structure and images in thumbnail grid view.

You can rate your images either using numbers or color tags. In the above image it has picked up the color rating I gave one image in LR. If you select the Catalog tab on the left hand menu, you can further refine your search parameters with selecting a specific rating or color tag. In the below example it has used the Red color tag to select images to view.

 

Also visible in the above image is the histogram (color graph below left) with camera settings above it for the selected image. The fine print at the bottom of the window has the name, file format, date/time taken, and file size information.

The full Manage mode window above, with directory tree/histogram/camera data on the left hand menu, and EXIF data for the selected image on the right hand pane, and all the images on display.

Other Image Viewing Options

ACDSee has two other image viewing options included. Photos mode and View mode

Photos mode opens with a splash screen explaining what it does.

It offers another way to sort and view your image files and has some granular control. You can get it down to a specific day quite easily and just see the images shot on that day. Probably very helpful for wedding or event photographers. Below is an example where it shows all the shooting days, with a blue bar that gives an idea of how many photos are stored under that day.

View mode is where you can see just a single image using the full screen size. You can zoom in to check the image quality using various zoom features. There is a floating Navigator panel you can activate and use that to ensure you are viewing the correct part of the image.  Similar to the Navigator in LR/PS.

There are some very basic editing tools available here, but better functionality is had in Develop mode.

4. Editing RAW Files

RAW image editing is done in Develop mode and it is laid out very similarly to LR. By default, the Editing tool panel is on the left but it can be moved.

Image with Edit Tool Panel on the left

It’s not immediately obvious, but the white section of the grey bar that ends with the triangle cut out of the bottom is the active slider. You move the light bar to the desired settings. Or type in a number or use the Up and Down arrows on the end.

There are 4 main tabs in the Tool Panel:

Tune – The usual tools for editing a RAW file, exposure, etc. Very similar to LR

Detail – Sharpening, Noise Reduction, and Skin Tuning

Geometry – Lens Correction, Cropping, Perspective adjustments

Repair – Heal/Clone and Red Eye adjustments

In Develop Mode with the Tune Panel open

In general, I found the sliders a bit fiddly to operate; it wasn’t smooth, but apparently it is easier to incrementally adjust sliders with a mouse wheel. My perception of the program is that its application of the settings is quite harsh, so careful use of the sliders is necessary.

While you can activate a second screen in Develop mode, the only purpose is to maintain a view of the unedited image for comparison.

The Tune tab also has some spot editing features—Develop Brush, Linear Gradient Tool and Radial Gradient Tool—the equivalent of Adjustment Brush, ND Grad, and Radial Tool in LR.

5. Advanced Editing With Layers

Edit mode gives most of the expected features you would find in Photoshop and other programs that offer layer/mask functionality. The Filmstrip is visible (similar to Bridge), although you can turn it off to gain the screen real estate back.

Edit Mode open with all the default settings and panels visible

Edit mode offers quite a few extra or useful features. The 2019 version also has an Adjustment layer for Color LUTs, which is a recent new feature brought into LR.

A new feature in the 2018 version was an Actions Menu—a range of preset creative edits you can apply with one click. The 2019 update to this allows you export and import actions as well.

Some of the actions have a really harsh effect like overdone HDR or similar, which was quite noticeable in the 2018 version. In the 2019 version they have toned down the effect in some of the actions, but not all of them. So it pays to pick and choose as it does depend on which action you choose as to what outcome you get. Also it applies it directly to the image so you can’t do it as a layer and then blend in, unless you duplicate the base layer and blend back which has its own issues.

One of the features that did impress me in both the 2018 and 2019 versions was how good a job the Heal tool did in tidying up spots and other issues. On the above image I have removed several spots and imperfections. On the right hand side, in the center of the flower, was a long black mark on a petal (near the small curled one), and that has been seamlessly removed.

An oddity also visible in the above image—in View mode I applied a LOMO preset and liked what it did, and further edited the image to mute the tones and lower the saturation.However, when you use the Navigator tool, as per above, it shows the original RAW file in its unedited state.

Finally I dragged some texture layers, (can be dragged from a second monitor into the Layer Palette), apply some blend modes, adjust the opacity, and soften areas with a mask to reach the final image.

New Features in 2019

Several new features have been included in the 2019 edition, but one key one is Face Recognition. A short video explains how to use it HERE.

I don’t shoot people/portraits generally but had a few tucked away to test. I could get the Face Recognition to function, however it didn’t automatically find all the other images and assign them correctly.  I suspect this is because I have all my images on a NAS and not in the usual directory. If I clicked on each image individually, it did recognise the face and the person.

General Comments

There are some things I find odd about how the program functions; three different ways to view the image can be a bit confusing. The second monitor view in Develop mode that only holds a copy of the unedited file for a comparison seems like a major waste of screen real estate.

Many new features were included in the 2018 version, and the ones assessed in this review of the latest version have been further enhanced and improved—I am guessing in response to user feedback.

This 2019 version adds a lot of nice new mature touches, and helpful splash screens to introduce you to different features.  More accessible help options is a vast improvement: there are links in the Help menu to a Support Community, a Facebook page, and a Twitter account.

Any new software program takes a bit of getting used to, but once you understand it, ADCSee Photo Studio Ultimate 2019 offers any beginner (and more experienced users) a compelling package. It has all the features you need for image management, RAW editing, and more advance editing in one place, with the advantage of a ‘pay once and it’s yours’ option instead of a subscription.  Although a subscription option is available, if desired.

At $ 149 USD for the single purchase perpetual licence, you get a LOT of capability all wrapped up in one software program.

Rating

8.5 out of 10

Disclaimer: ACDSee is a paid partner of dPS

 

 

 

The post ACDSee Photo Studio Ultimate 2019 Review: Face Detection and Recognition appeared first on Digital Photography School.


Digital Photography School

 
Comments Off on ACDSee Photo Studio Ultimate 2019 Review: Face Detection and Recognition

Posted in Photography

 

Video: Tamron 28-75mm F2.8 FE face and eye-detect autofocus test

18 May

The Tamron 28-75mm F2.8 Di III RXD for Sony FE cameras made a big splash when it was first teased back in February. People were intrigued by its small size and the new Rapid eXtra-silent stepping drive (RXD) AF motor; it even came up in our CP+ interview with Tamron, and we got to see the lens in person at the show.

The lens isn’t going to be officially available for another week; however, photographer David Oastler was able to get his hands on a copy and, while he wasn’t allowed to take photos with it, he was allowed to put it through it’s autofocus paces to see how that RXD motor holds up.

What Oastler really wanted to see is how well the Eye and Face-detect autofocus from the Sony FE body would perform through this third-party lens that was, ostensibly, designed from the ground up to work on this full-frame mirrorless system. While the video isn’t the best quality (a bit of glare) you can still see, and Oastler tells you, that the lens performs exceptionally well. In fact, Oastler goes so far as to say he noticed no performance difference between the Tamron and his own Sony-native lenses.

Tamron promised as much when it released the lens, calling it “quiet, precise, and exceedingly quiet.” But it’s nice to see a real-world test confirm these claims.

We’ll be trying to get our hands on a Tamron 28-75mm F2.8 Di III RXD as soon as humanly possible for our own in-depth testing. But in the meantime, if you’re interested in picking up this $ 800 USD lens when it ships at the end of next week and you want to see how its AF motor performs IRL, check out Oastler’s video at the top.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Video: Tamron 28-75mm F2.8 FE face and eye-detect autofocus test

Posted in Uncategorized

 
« Older Entries