Hello, My project is about building a model to detect emotions. I got a dataset called affwild2 with images tagged with the following emotions “Neutral”, “Anger”, “Disgust”, “Fear”, “Happiness”, “Sadness”, “Surprise”, “Other”. The training dataset has this number of images: Anger → 16573 Dislike…

[image] Rodrigo_Cotarelo: The accuracy reaches 95% but in the validation dataset, it does not exceed 40%. I will assume the first number is training accuracy. If you haven’t studied up on the concepts of bias and variance, you should do. High training accuracy but low validation accuracy ind…

Thanks for answering. I continue improving the model and i saw the Hyperparameters course. These are my actuals results: [image] I think now i am facing a data problem. The variance between the traning a validation data It seems to be due to the lack of data on some emotions. It just so happens t…

Can I ask you one thing, is fear and disgust in the other group of dataset?

Hello @Rodrigo_Cotarelo I don’t know if this will help you, came across a similar project but not using vgg but it gives quite a detail attention on subject dependent and subject independent review, probably it will help you grab some more pointers in the direction you are looking for. Emotional …

also go through these screenshots for understanding bias and variance issue [Screenshot 1945-01-19 at 9.40.56 PM] [Screenshot 1945-01-19 at 9.43.38 PM] [Screenshot 1945-01-22 at 12.02.06 PM] [Screenshot 1945-01-22 at 12.04.59 PM] Regards DP

Hi DP, Thanks for answering! I appreciate it. By other group of dataset do you mean if the photos of fear and disgust are from other dataset?

Yes, also I am not sure if the dataset images dimensions, if the images are normalised? before using this dataset to the vgg model

I am using the aff-wild2 dataset. The Aff-Wild2 is annotated in a per frame basis for the seven basic expressions (i.e., happiness, surprise, anger, disgust, fear, sadness and the neutral state), twelve action units (AUs 1,2,4,6,7,10,12,15,23,24,25, 26) and valence and arousal. In total Aff-Wild2 …

Hello @Rodrigo_Cotarelo I don’t know if it is allowed, could you share some images only few!! [image] Rodrigo_Cotarelo: The Aff-Wild2 is annotated in a per frame basis for the seven basic expressions (i.e., happiness, surprise, anger, disgust, fear, sadness and the neutral state), twelve actio…

Emotion Recognition Project

AI Discussions

Deepti_Prasad March 26, 2024, 3:07pm 7

also go through these screenshots for understanding bias and variance issue

Screenshot 1945-01-22 at 12.02.06 PM

Regards
DP

I made a deep learning model and after 100 epochs the graph of validation accuracy is having too many oscillations with every epoch. How do we normalize the validation accuracy. Although the validation loss is satisfactory.

Topic		Replies	Views
Improving validation accuracy of an image sentiment identification system AI Discussions ai-discussions , project	4	150	April 21, 2024
CNN - Facial Emotion recognition Convolutional Neural Networks week-module-2 , ai-discussions , project	2	89	July 9, 2025
Why am I getting the same accuracy on validation data? - Python - Computer Vision - Deep Learning AI Discussions ai-discussions	3	113	December 28, 2024
Test Accuracy Higher than Train accuracy? Improving Deep Neural Networks: Hyperparameter tun coursera-platform	42	2938	August 9, 2024
Personal image recognition exercise, loss/error rate to large Advanced Learning Algorithms week-module-2	13	702	September 5, 2022

Emotion Recognition Project

Related topics