When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
I'm working on a class project for a deep learning course I'm taking, centered on building an image classifier. If any of you might be willing to take a moment to submit a picture of your 9 Inch or Dana 60 rear end (taken from the back of the truck or as it would be viewed from the back of the truck if not installed, for the sake of consistency), I'd really appreciate the help.
I just uploaded three photos of a nine inch. Lighting isn’t very good but I figure machine learning needs to adapt to lighting issues.
i can probably get a few Dana 60 shots later off the f250 frame that’s sitting in my yard.
Awesome, thanks so much! Less-than-perfect pictures are great; they should hopefully feed into a more robust model in the end. The motivating question for the project, which I didn't really bring up in the original post, was whether image-label pairs generated by subject matter experts could form a good dataset to train a model for use in niche topic areas: thinking that, for example, you might never see a Captcha asking to pick Dana 60 versus 9 Inch rear ends, and nobody probably wants to put in the time scraping the internet for pictures of each, but a model that can pick between them might be useful in a setting where people constantly come in asking "what is X?".
For a monkey wrench, I have a Dana 44 rear end from my '66, sitting in my yard
A very fair point; I'm a little worried the generated set of images might come out a bit small, so I'm absolutely cheating by arbitrarily keeping it to 2 classes of image