Machine vision expert to provide guidance on neural network algorithm for image classification

We are training a convolutional neural network to perform image classification of food images. We have a small-ish dataset of 55,000 images across 600+ labels. We have tried these 3 approaches for building our model. Attempts at fine-tuning a pre-trained model or using bottleneck features have led to overfitting during training and poor model performance. Our initial attempt at building our own network from scratch yields promising results in terms of appropriate accuracy and loss curves for training and validation sets.

We are looking for someone who:
- Can provide guidance and suggestions for building a model that performs better given that getting more data is not an option
- Has ability to have a discussion around our existing Keras models
About the recuiter
Member since May 20, 2018
Ankur Patidar
from Lazio, Italy

---

---

Looking for help? Checkout our video tutorial
How to search and apply for jobs

How to apply? Do you have more questions about the Job?
See frequently asked questions