VSaaS.ai - Plataforma de Video Analítica con IAvsaas.ai
AI Tool

Custom Training

Create your own custom object detection models without code.

The Custom Training tool allows users to create fully customized object detection models from the VSaaS.ai web interface, without any machine learning knowledge. The complete workflow includes uploading images from IP cameras, labeling objects with bounding boxes, defining custom classes, configuring training hyperparameters, and launching the process with one click. Ideal for detecting industry-specific objects not covered by the 46 standard analytics.

Custom Training1📷 100+23MediumEpochs: 50Batch: 880/20468%5Train custom AI models with your own data

Core Features

Discover the capabilities that make this tool indispensable for your operation.

Batch Image Upload

Drag and upload images directly from your IP cameras organized by batches with date and description.

Visual Labeling

Annotation tool with bounding boxes: draw, select, and classify objects directly on images.

Custom Classes

Define the object classes you need to detect (e.g., empty-slot, product-A, helmet, PPE) with unlimited categories.

Auto-Labeling

Automatic mode that pre-labels images using existing models to accelerate the annotation process.

Training Configuration

Adjust model size (S/M/L), epochs, batch size, image size, and train/val percentage.

One-Click Training

Launch custom model training with a single click and monitor progress in real time.

Key Benefits

Create custom detection models without machine learning expertise.
Detect industry-specific objects not covered by standard analytics.
Complete workflow from image capture to production model on the same platform.
Reduces custom model development costs by eliminating the need for data science teams.
Fast iteration: upload more images, re-label, and re-train to improve accuracy.
Labeled datasets become strategic company assets.

Technical Specs

Supported taskObject detection (DET)
Model sizesS (small), M (medium), L (large)
Image size640px (configurable)
EpochsConfigurable (default: 50)
Batch sizeConfigurable (default: 8)
Train/Val split80/20 (configurable)
Image formatsJPG, PNG from IP cameras

Use Cases

Real-world scenarios where this tool delivers value for your organization.

1

Retail — Empty Shelf Detection

Store chain trains a model to detect 'empty-slots' on product shelves, automating visual inventory control with existing cameras.

2

Industry — Custom PPE

Industrial plant trains a model to detect their specific protective equipment (logo helmets, corporate color vests) that generic models don't recognize.

3

Agriculture — Pest Detection

Agricultural company trains a model to identify specific types of pests or diseases in crops using field camera images.

4

Logistics — Packages & Containers

Distribution center trains a model to classify package types, detect container damage, and verify shipping labels.

Ready to implement Custom Training?

Schedule a personalized demo and discover how this tool can transform your video analytics operation.