Hello,
at the end of second ungraded lab, there is a note:
If your training reached 80% accuracy, you may notice from the images above that the presence of eyes and nose play a big part in determining a dog, while whiskers and a colar mostly point to a cat.
Why do you say this? How can you notice? There are also cats with eyes and nose…
Thanks
Hi @Dennis_Sinitsky
The text is just saying the same thing as the images. The resulting mapping generated by CAM differs for cats and dogs, which is what the text says.
Keep learning!
Hello @Dennis_Sinitsky
Eyes and nose are considered facial landmarks in detection algorithm, hence it is basically not only about cats and dogs, but facial features related to cats and dogs holds significance.
Regards
DP
For dogs the differentiating features seem to be eyes and nose and for cats seem to be whiskers and collars, these are how they mainly differ and what the model is able to learn!
Thank you for your replies. In case of whiskers for cats – it does seem that cats are emphasized. And clearly cats have pointier ears, although some ear pixels for dogs also come out. But as I looked at the images, it seemed that both cats and dogs have nose and especially eye pixels light up; not sure why it is claimed that it tends to emphasize dogs. May be I do not completely understand the meaning of “play a big part in determining…”.