Tasvirlarni klassifikatsiya qilishda xususiyatlarni birlashtirish usuli orqali model aniqligini oshirish
Keywords:
Global average pooling, softmax funksiyasi, neyron tarmoq, Efficient NetB0Abstract
Ushbu maqolada tasvirlarni klassifikatsiya qilishda model aniqligini
oshirish uchun xususiyatlarni birlashtirish usullari tahlil qilinadi. EfficientNetB0
modelining ishlatilishi, xususan, Global Average Pooling (GAP) va Softmax funksiyalari
orqali modelning sinflar orasidagi farqlash qobiliyati oshirilgani ko„rsatilgan. GAP usuli
fazoviy o„lchamlarni qisqartirish orqali har bir kanal uchun o„rtacha qiymat hisoblash
imkonini beradi va maxsus tortishish koeffitsientlariga ehtiyoj sezmasdan, samaradorlikni
oshiradi. Softmax esa chiqish qiymatlarini ehtimollik sifatida ifodalash imkonini beradi. Bu
maqola, ayniqsa, tasvir klassifikatsiyasiga qaratilgan konvolyutsion neyron tarmoqlarni
samarali ishlatishda foydali bo„lishi mumkin.
References
Li, M. et al. (2020). "Global Average Pooling in Convolutional Neural Networks."
Howard, A. et al. (2019). "EfficientNet: Rethinking Model Scaling for Convolutional
Neural Networks."
Goodfellow, I., Bengio, Y., Courville, A. (2016). "Deep Learning." MIT Press.
Simonyan, K., Zisserman, A. (2015). "Very Deep Convolutional Networks for Large
Scale Image Recognition."
Krizhevsky, A., Sutskever, I., Hinton, G.E. (2012). "ImageNet Classification with Deep
Convolutional Neural Networks."
He, K. et al. (2016). "Deep Residual Learning for Image Recognition."
Chollet, F. (2017). "Xception: Deep Learning with Depthwise Separable Convolutions."
Szegedy, C. et al. (2015). "Going Deeper with Convolutions."
Zeiler, M.D, Fergus, R (2014) "Visualizing and Understanding Convolutional Networks."
Girshick, R. (2015). "Fast R-CNN."
