Multiple object detection in Lesson 2

tedmasterweb · October 21, 2021, 11:40am

Hi,

We’re still learning with fast.ai and really enjoying it. It makes so many things so much faster. However, we’re looking for a way to optimize some parts of the computer vision stuff.

Referring to the ‘lesson2’ notebook with bears… What happens if you have two or more objects on the image belonging to two or more different classes: namely an image with a black bear and a grizzly bear. Which of the two objects will the net recognize? Can we make the net recognize and identify both the objects? to what folder would we need to insert that picture in that case? How do we deal with this scenario?

TIA

Ted Stresen-Reuter

muellerzr · October 21, 2021, 1:38pm

That would be Multi-Label Classification, which is gone over later in the course

github.com

fastai/fastbook/blob/master/06_multicat.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#hide\n",
    "!pip install -Uqq fastbook\n",
    "import fastbook\n",
    "fastbook.setup_book()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [

This file has been truncated. show original

tedmasterweb · October 21, 2021, 2:41pm

Thank you very much. That’s exactly what we were hoping to find.

Ted