Technology that can accurately identify animals can help reunite owners with lost animals, help farmers monitor livestock, and help researchers study wildlife. Historically, microchips have been the most popular method of animal identification in this regard. However, implanting the chip requires invasive surgery. They cannot be read without specialized equipment, and thieves can extract the microchips. Another method is DNA analysis, which is accurate but also very expensive and time-consuming.
Animal facial recognition (and sometimes not just faces) powered by computer vision solutions can serve as a viable alternative to the above methods. Although it has its shortcomings, this technology can show a high level of accuracy in certain situations. So, how does animal facial recognition work? What are the challenges hindering the advancement of this technology?
Animal facial recognition solutions in general There are three main steps:
Image capture: Take photos of animals with a high-resolution camera. Some algorithms only work on predefined poses, so images that meet these criteria must be selected.
Feature Extraction: Evaluate the animal’s biometric data for suitability and perform preprocessing if needed. The algorithm then extracts the feature set required for recognition.
Matching: The extracted features are mathematically represented and matched with other images. For example, if we are looking for a dog in a lost pet database, we match the dog's unique characteristics to all the animals in the database.
Several ways to perform matching. One method is to use algorithms such as KNN and DBSCAN for clustering to obtain a set of images that are highly close to our target image, and the user can manually select the most suitable image. Alternatively, probabilistic methods can be used to express the final results as confidence levels.
Losing a pet is heartbreaking for its owner. And according to statistics, this is much more common than people think. In the United States, one in three household dogs and cats has gone missing at some point in their lives, and 80% of them have never been recovered. There are several tools based on pet facial recognition that can help owners find their lost friends.
ForPaws: This animal facial recognition solution identifies dogs based on the tip of their nose, skin color, and fur type. Animal owners are asked to upload at least three photos to create a "personal profile" of their animal. Currently, the program can identify 130 dog breeds with an accuracy of 90%.
PiP: This animal identification company has developed an app that allows animal owners to register and upload photos of their animals. The system analyzes their unique facial features. PiP claims it can identify every lost cat and dog if the owner provides more information, such as gender, size and weight.
Anyone who finds a lost pet can also use the app to find the owner. PiP’s solution also continuously scans social media for pet posts and sends missing pet alerts to residents in relevant communities.
Love Lost: Love Lost by Petco is another app that helps pet owners and pet shelters. Owners are advised to create profiles of their pets so that when a pet goes missing, the software can begin matching the animal's biometric information to new shelter members and other candidate pets.
Sometimes, it makes sense to train an algorithm to recognize specific animals. For example, animal owners could benefit from a system that accurately identifies their animal and takes appropriate action, such as sending an alarm or opening a door to let the animal in. Arkaitz Garro, a front-end engineer at WeTransfer, developed an animal facial recognition solution that can identify a neighbor's cat and send Garro an alert when the cat shows up at the door.
To capture photos of the cat, Garro used a small camera and a Raspberry Pi with motion detection software. When an animal approaches the camera, a photo is taken and sent to the AWS recognition platform for comparison with other photos of the cat uploaded by Garro. If there is a match, the engineer will be notified.
Microsoft has also developed an Internet of Things (IoT) device that can perform animal recognition and can be connected to the pet portal. Once it recognizes that it's your pet, the device opens the door and lets it inside.
In addition to identifying household animals, facial recognition algorithms can also be used to identify other species. A study published in the Journal of Marine Mammal Science looked at a set of characteristics needed to identify dolphins. Researchers tracked and photographed 150 bottlenose dolphins over 12 years. The research team wanted to evaluate the idea of using a dolphin's face and dorsal fin for identification throughout its life.
Of the 150 experimental subjects, only 31 dolphins had complete profiles (that is, clear photos of the left and right sides of the face and dorsal fins). The study relied on human expert opinion and statistical methods to detect similarities between different images of the same dolphin.
The experimental results show that dolphin facial features remain consistent over time and can be used for identification purposes. The ability to identify pups even when they are adults has greatly facilitated the study of dolphins.
Identifying farm animals can be a challenging process. With pigs, it's more difficult because all pigs look the same. But cows are a bit special. They are black and white and have different shapes. However, when it comes to cows, another challenge arises – where to install the cameras. Cows are curious animals and will notice even the smallest changes in their surroundings. They often try to lick or otherwise interact with the camera.
But building a system that can identify individual cows would help farmers tremendously. This solution matches an animal's health and dietary patterns to the animal's identity. Enhanced with artificial intelligence, it will be able to detect any signs of disease and abnormal behavior and notify farmers in case of emergency.
The core algorithm platform of Beijing Xiangchuang Technology has realized data collection and facial recognition of pigs, cattle, sheep, donkeys and other livestock, and has accumulated more than tens of millions of livestock facial data. It not only helps farmers carry out refined breeding management, but also assists banks, insurance and other financial institutions to establish risk assessment and early warning systems when conducting business in the breeding industry.
Facial recognition technology for animals lags far behind current fairly advanced facial recognition technology for humans. Researchers began experimenting with animal facial recognition about four years ago, but the accuracy of common techniques is still quite low. On the other hand, solutions with a specific purpose, such as identifying a specific animal, can be accurate.
Companies that want to implement animal facial recognition solutions need to consider three main challenges:
Scientists have specified a feature vector, Can be used for unique face recognition. However, the same approach doesn't work for animals because we don't know which features we need to use and how to interpret them. For example, when working with people, scientists can use variational autoencoders (VAE) architectures to extract features from faces. In this method, a photo of a person is compressed into a vector containing the desired features, such as skin tone and facial expression.
When it comes to animal facial recognition, there are currently no reliable feature vectors. Solving the challenge of a reliable eigenvector will greatly advance research in this area.
An open source example in this regard is DogFaceNet, which is a deep learning-based implementation of dog recognition. It uses the dog's eyes and nose as a feature set. This solution works reasonably well if the overall goal is to distinguish dog breeds, but when it comes to distinguishing individual animals it performs rather poorly.
Another example is to use the local binary pattern histogram (LBPH) algorithm, which converts images into pixels and operates by comparing the pixel values of different images . This method depends on the posture of the animal, which makes it sensitive to changes in posture.
For humans, it's easy to assume a specific pose and sit still. However, things get more complicated when we try to get a cat or dog to hold still in a specific position.
For training to be effective, the data must be diverse and cover all tasks the algorithm is expected to perform. For example, if the algorithm is supposed to identify different dog breeds, then the data set should adequately cover all the breeds captured from different angles and be labeled appropriately. There are several things that can go wrong here. For example, someone might submit a picture of a mixed breed, and someone might label their picture incorrectly and assign the wrong breed name. To avoid such problems, experts must review all photos in the dataset one by one to verify the legitimacy of the images and the accuracy of the labels.
Progress in the field of animal facial recognition has been hampered because researchers still cannot pinpoint the optimal combination of features that can be used to accurately identify animals at scale. Still, there are some successful applications that operate on limited data, such as identifying a specific animal or a small group of domestic or wild animals.
If you are building your own animal facial recognition system, keep in mind that animals are uncooperative biometric users. Some will insist on licking the camera, and some will refuse to stand up for a photo.
The above is the detailed content of What is animal facial recognition technology used for?. For more information, please follow other related articles on the PHP Chinese website!