I had an ESP32CAM available with me for a long time and always wanted to be able to detect any object of my choice using it.
Luckily I found edge impulse and after bouncing off a lot of tutorials I was able to train my model and achieve what I wanted. To make it easy for others I have documented my complete tutorial here called
ESP32CAM Object Detection
I found that training the model by using the images taken from the ESP32 cam and clicking images in an isometric angle gave the best results. Adding this information to the forum so as to help people find it in the future if needed.
Cheers!!