Object Detection - cocoSsd (Userscript)
Mode Homepage & Docs
In Diffgram
Search cocoSsd
Video Example
Code
if (!this.model) {
this.model = await cocoSsd.load({
base: 'mobilenet_v2'
})
} else {
console.log("used cached model")
}
const warm_up = await this.model.detect(
diffgram.get_new_canvas())
let canvas = diffgram.get_new_canvas()
const predictions = await this.model.detect(
canvas)
for (let prediction of predictions){
console.log(prediction)
if(prediction.bbox[2] <50) {continue}
diffgram.create_box(
prediction.bbox[0],
prediction.bbox[1],
prediction.bbox[2] + prediction.bbox[0],
prediction.bbox[3] + prediction.bbox[1])
}
NPM Packages
https://cdn.jsdelivr.net/npm/@tensorflow/tfjs
https://docs.opencv.org/master/opencv.js
https://cdn.jsdelivr.net/npm/@tensorflow-models/coco-ssd
Open cv requirement is a workaround due to this issue.
Example Image
This uses mobilenet_v2
Updated over 3 years ago