Chooch Solutions for the Public Sector

All Chooch Enterprise Licenses include access to all functions of the Chooch platform for each named user, whether Cloud or Edge. Chooch provides users with full control of the visual AI lifecycle: Dataset generation, Annotation and Labeling, Model Training and Testing, Edge AI device creation and management, Inference Engine Export, API integration, and Live Video Stream Management.

  • User Interface
    • Cloud Dashboard, Edge AI Device Control Panel
  • Pretrained Models
    • Public Object Models, Public Face Models, Public Image Models
  • Data Generation and Management Tools
    • Existing Dataset Upload, Raw Data (Video or Still Image) Upload, Object Annotation (Still Image or Video), Smart Annotation, Data Augmentation, 2D/3D Synthetic Data Generation
  • Model Training and Testing
    • Training and Testing of Object Detection Models, Image Classification Models, Facial Recognition Models, Manual Model Testing, Automatic F1 Score Generation, and Dataset generation for training and testing
  • Edge AI Device Creation and Data Management
    • Device Creation, AI Inference Engine Export, Edge AI Control Panel, MQTT Broker, Cloud API Access, Edge Device API Access
  • Live Video Stream Management
    • On-Premises and Edge AI Processing
  • Dashboard

    The User Interface for the Chooch computer vision software platform. Non-technical end-users can access all functions of the software without coding.

  • Dataset Generation

    Chooch Image and Video Annotation tools enable users to annotate a single frame. The object tracking layer does the rest - tracking the object from frame to frame, autonomatically forwarding the user-specified object label. Datasets can be generated from raw video, CAD files, or from existing COCO JSON datasets. Smart Annotation allows rapid re-labeling of objects to meet changing requirements. Data Augmentation options allows users to modify raw data to replicate different sensor types and atmospheric conditions.

  • Analytic Model Training and Testing

    Users select and combine annotated datasets to train models. Model training generates an F1 score, the measure of a model's accuracy on a dataset. Users can accept the model or retrain it to the desired accuracy threshold.

  • Edge Device Management

    Users can choose the type of Edge processing device - CPU or GPU. Drivers are automatically compiled and downloaded to the device. Once the device is created, users can add models and IP addresses of video feeds for analytics at the Edge. Use of MQTT Broker is also an option to enable autonomous notifications, alerts, and machine to machine response (IoT).

  • Inference Engine

    The Chooch inference engine is the brain of the AI software. Chooch’s ability to generate a “Dense Classification” - detecting, recognizing, geolocating, and reporting all of the objects within a frame, is covered under patent. Chooch generates a

    Contextual Perception, which means that not only does Chooch ‘see’ all of the objects, it understands their collective relationship. The Chooch inference engine merges Image, Object, Face, Text, Action, Anomaly and Change Detection, Counting, and Unknown Object detection. It is also capable of In-Object detection for the highest degree of analysis.

  • On-Premises/Self-Hosted

    The On-Premises (aka Self-Hosted) solution is for customers requiring the highest degree of data management and security. Designed for installation in the air-gapped environment, the on-premises solution includes the full-functionality of Chooch software while reducing the risk of data spillage, latency, or data stream interruption.