53503 - پروژه R
سه سال پیش منتشر شده
تعداد بازدید: 126
کد پروژه: 398034
شرح پروژه
سلام لطفا ویسهایی ک ضمیمه کردم رو گوش کنین
اینم توضیح کار هست "
For this project you will use the Cats and Dogs data set: 99 images of Cats and 99 images of Dogs. Each image is 64 by 64 pixels so it's pretty low resolution. The data set is of dimension 198 by 4096 (64 times 64), i.e. raster scans of the images. Label 0 denotes a cat and Label 1 a dog. The csv files containing the images and labels are on Canvas.
For the project, we will treat pixel values as features, ignoring the fact that these are images. What I mean, you are not expected to use image analysis tools etc. You can use the spatial (co)localization of the pixels for interpretation of results or to guide your methods. Visualizing e.g. feature importance or selection or mislabeled observations etc can be done using image representations for example.
Theme 2b
The images are 64 by 64. You can view the images as built from 16 pixel blocks of size 16 by 16 - corner blocks, a next to corner blocks or interior blocks. Your task is to use a statistical method to choose one of these 16 by 16 blocks (the same block for all images) as input into the classification methods.
Is there a block of that size that results in good classification performance?
این پروژه شامل 5 فایل مهم است، لطفا قبل از ارسال پیشنهاد حتما نسبت به بررسی این فایل اقدام فرمایید.
مهارت ها و تخصص های مورد نیاز
مهلت برای انجام
2روز
وضعیت مناقصه
بسته
درباره کارفرما
عضویت چهار سال پیش
