Advanced Driver Vision Assistance with Near Collision Estimation System

API for object detection, depth estimation, and distance prediction using computer vision models

Overview

This API provides access to advanced computer vision models for real-time image processing. It leverages:

The API supports both HTTP and WebSocket protocols:

POST /api/predict

Process a single image via HTTP request

WS /ws/predict

Stream images for real-time processing via WebSocket

HTTP API
WebSocket API
Try it

HTTP API Reference

POST /api/predict

Process a single image for object detection, depth estimation, and distance prediction.

Request

Content-Type: multipart/form-data

Parameter Type Required Description
file File Yes The image file to process (JPEG, PNG)

Request Example

# Python example using requests import requests url = "http://localhost:8000/api/predict" files = {"file": open("image.jpg", "rb")} response = requests.post(url, files=files) data = response.json() print(data)

Response

Returns a JSON object containing:

Field Type Description
objects Array Array of detected objects with their properties
objects[].class String Class of the detected object (e.g., 'car', 'person')
objects[].distance_estimated Number Estimated distance of the object
objects[].features Object Features used for prediction (bounding box, depth information)
frame_id Number ID of the processed frame (0 for HTTP requests)
timings Object Processing time metrics for each step

Response Example

{
  "objects": [
    {
      "class": "car",
      "distance_estimated": 15.42,
      "features": {
        "xmin": 120.5,
        "ymin": 230.8,
        "xmax": 350.2,
        "ymax": 480.3,
        "mean_depth": 0.75,
        "depth_mean_trim": 0.72,
        "depth_median": 0.71,
        "width": 229.7,
        "height": 249.5
      }
    },
    {
      "class": "person",
      "distance_estimated": 8.76,
      "features": {
        "xmin": 450.1,
        "ymin": 200.4,
        "xmax": 510.8,
        "ymax": 380.2,
        "mean_depth": 0.58,
        "depth_mean_trim": 0.56,
        "depth_median": 0.55,
        "width": 60.7,
        "height": 179.8
      }
    }
  ],
  "frame_id": 0,
  "timings": {
    "decode_time": 0.015,
    "models_time": 0.452,
    "process_time": 0.063,
    "json_time": 0.021,
    "total_time": 0.551
  }
}

HTTP Status Codes

Status Code Description
200 OK - Request was successful
400 Bad Request - Empty file or invalid format
500 Internal Server Error - Processing error

WebSocket API Reference

WebSocket /ws/predict

Stream images for real-time processing and get instant results. Ideal for video feeds and applications requiring continuous processing.

Note: WebSocket offers better performance for real-time applications. Use this endpoint for processing video feeds or when you need to process multiple images in rapid succession.

Connection

# JavaScript example const socket = new WebSocket('ws://localhost:8000/ws/predict'); socket.onopen = function(e) { console.log('Connection established'); }; socket.onmessage = function(event) { const response = JSON.parse(event.data); console.log('Received:', response); }; socket.onclose = function(event) { console.log('Connection closed'); };

Sending Images

Send binary image data directly over the WebSocket connection:

// JavaScript example: Sending an image from canvas or file function sendImageFromCanvas(canvas) { canvas.toBlob(function(blob) { const reader = new FileReader(); reader.onload = function() { socket.send(reader.result); }; reader.readAsArrayBuffer(blob); }, 'image/jpeg'); } // Or from input file fileInput.onchange = function() { const file = this.files[0]; const reader = new FileReader(); reader.onload = function() { socket.send(reader.result); }; reader.readAsArrayBuffer(file); };

Response Format

The WebSocket API returns the same JSON structure as the HTTP API, with incrementing frame_id values.

Response Example

{
  "objects": [
    {
      "class": "car",
      "distance_estimated": 14.86,
      "features": {
        "xmin": 125.3,
        "ymin": 235.1,
        "xmax": 355.7,
        "ymax": 485.9,
        "mean_depth": 0.77,
        "depth_mean_trim": 0.74,
        "depth_median": 0.73,
        "width": 230.4,
        "height": 250.8
      }
    }
  ],
  "frame_id": 42,
  "timings": {
    "decode_time": 0.014,
    "models_time": 0.445,
    "process_time": 0.061,
    "json_time": 0.020,
    "total_time": 0.540
  }
}

Try The API

You can test the API directly using the interactive Swagger UI below:

Simple WebSocket Client

Upload an image to test the WebSocket endpoint:

Status: Disconnected

Last Response: