AI models are now available for a suite of use cases – deployed to the edge or the cloud. Examples include Mask Detection and Compliance and Social Distancing Assurance.
Chooch AI also trains and develops flexible, custom AI models to extend standard models. Customer deployments include quality control for online media, facial authentication for touchless verification, safety equipment compliance and remote oil and gas well monitoring.
AI models are packaged with applications and are installed remotely on leading edge devices to provide fast, accurate edge AI in minutes – or AI models can be accessed via API in the cloud.
Chooch Edge AI is an inference engine generated by Chooch Cloud AI that can be remotely deployed to industry leading NVIDIA Jetson processors. Chooch Edge AI deployments are remotely managed and can be updated from the Chooch AI Dashboard in the cloud.
Currently, Chooch Edge AI can deploy up to 8 models and 8,000 classes for robust Visual AI on a single NVIDIA Jetson Nano edge device. Devices running Chooch Edge AI are assigned unique IDs and can interpret data from up video streams from multiple cameras simultaneously. Chooch AI inference engines are very fast, generating responses under 0.02 seconds, and processing 10 simultaneous calls per second.
The Chooch AI Dashboard is the gateway to the 200,000 public visual AI classes. Use the Chooch Dashboard for AI training, model deployment, inferencing, edge device management and the API integration with information systems. With the dashboard, non-engineers can conduct object, image, and facial authentication training.
When you create a free account, you will receive an API key for integration and gain access to Chooch AI models, frameworks, and public classes. You will also be provided API documentation, videos, reference materials for end-user self-training as well as a link to customer support.
Facial authentication works very similarly to identifying objects in photos and videos with metadata, but because it involves human identity, use of facial authentication will always be consent-based.
Chooch AI can be trained with as few as five images or a quick video. Chooch AI matches a face in a photo or video using 512 biometric features
And by integrating Liveness Detection into our authentication process, Chooch AI cannot be spoofed with a “presentation attack” consisting of a video of a person’s face, a mask, or other hacks. Facial authentication has many use cases, ranging from account access to digital signature services.
The team at Chooch AI works in partnership with organizations to train AI models.
We can support your team by labelling and annotating images and video. Alternatively, our team can help develop models specific to your use case, annotate your data set, and train our AI to identify the objects, actions, images for any visual task.
Examples of successful deployments fully trained by the team at Chooch include operating room procedure imagery, images of cars for quality control, images of cell nuclei, and geospatial images of intact and damaged buildings.
Our flexible annotation tool can be used to train on still images, and the video annotation process can generate over 1,000 annotated images per minute. Learn more about AI training and annotation.
Please contact us to discuss your visual AI needs. After you describe the problem you are trying to solve, we will request images and/or video in order to create a new set of trained perceptions to execute your project. Once trained, your proprietary perception library will be enabled to integrate by API with external imagery sources.
Chooch AI is easily integrated into video management systems and can send still images and metadata alerts to any type of dashboard – from medical imaging systems to public safety networks.
Chooch AI supports public safety by detecting the absence of mandated safety equipment such as masks, and the signs of illness such as coughs or fever in public places. This model does not include any facial recognition features.
Chooch AI also supports x-ray analysis to detect lung injuries. Advanced medical imagery analysis both enables the detection of pneumonia and other conditions that lead to Acute Respiratory Distress Syndrome (ARDS) and can accelerate research.