I need to evaluate and compare object detection algorithms (YOLO and SSD) trained on COCO dataset. The metrics I have to use are precision, recall and mAP.
What I don't understand is how I can do it if my test dataset, which I have to film by myself, is not labeled. There are no ground truth boxes on my footage so I can not calculate IoU. Can I just do it "by eye", meaning that if it is obvious that IoU is > 0.5 I consider it a true positive?
I tried implementing YOLOv3 and SSD with pre trained weights. I manage to detect objects quite good on my test dataset, but I need to create a report with all the metrics mentioned. I can tune "confidence threshold", but that is not it.
What would be the best thing to do to finish my task without having to label my test data?
Related
Data augmentation is surely a great regularization method, and it improves my accuracy on the unseen test set. However, I do not understand why it reduces the convergence speed of the network? I know each epoch takes a longer time to train since image transformations are applied on the fly. But why does it affect the convergence? For my current setup, the network hits a 100% training accuracy after 5 epochs without data augmentation (and clearly overfits) - with data augmentation, it takes 23 epochs to hit 95% training accuracy and never seems to hit 100%.
Any links to research papers or comments on the reasonings behind this?
I guess you are evaluating accuracy on the train set, right? And it is a mistake...
Without augmentation your network simply overfits. You have a predefined number of images, for instance, 1000, and your network during training can easily memorize dataset labels. And you are evaluating the model on the fixed (not augmented) dataset.
When you are training your network with data augmentation, basically, you are training a model on a dataset of infinite size. You are doing augmentation on the fly, which means that the model "sees" new images every time, and it cannot memorize them perfectly with 100% accuracy. And you are evaluating the model on the augmented (infinite) dataset.
When you train your model with and without augmentation, you evaluate it on the different datasets, so it is not correct to compare their accuracy.
Piece of advice:
Do not look at train set accuracy, it is simply misleading when you use augmentations. Instead - evaluate your model on the test set (or validation set), which is not augmented. By doing this - you'll see the real accuracy increase for your model.
P.S. If you want to find out more about image augmentaitons, I really recommend you to check this guide - https://notrocketscience.blog/complete-guide-to-data-augmentation-for-computer-vision/
I am trying to train a CNN model for a regression problem, after that, I categorize predicted labels into 4 classes and check some accuracy metrics. In confusion matrix accuracy of class 2,3 are around 54% and accuracy of class 1,4 are more than 90%. labels are between 0-100 and classes are 1: 0-45,2: 45-60, 3:60-70, 4:70-100. I do not know where the problem comes from Is it because of the distribution of labels in the training set and what is the solution! Regards...
I attached the plot in the following link.
Training set target distribution
It's not a good idea to create classes that way. Giving to some classes a smaller window of values (i.e. you predict 2 for 15 values and 1 for 45 values), it is intrinsically more difficult for your model to predict class 2, and the best thing the model will learn during training will be to avoid class 2 as much as possible.
You may confirm this having a look at False Negatives for classes 2 and 3, if they are too many, it might be due to this.
The best thing to do would be categorizing your output space in equal portions, and trusting your model will learn which classes are less frequent, without trying to force that proportion by yourself.
If you don't have good results, it means you have to improve your model in other ways, maybe using data augmentation to get a uniform distribution of training samples may help.
If this doesn't sound convincing for you, try to have a look at this paper:
https://papers.nips.cc/paper/95-alvinn-an-autonomous-land-vehicle-in-a-neural-network.pdf
In end-to-end models for autonomous driving, neural networks have to predict classes indicating the steering angle. The distribution of these values is highly imbalanced as most of the time the car is going straight. Despite this, the best models do not discriminate against some classes to adapt to data distribution.
Good luck!
I am training a CNN model(made using Keras). Input image data has around 10200 images. There are 120 classes to be classified. Plotting the data frequency, I can see that sample data for every class is more or less uniform in terms of distribution.
Problem I am facing is loss plot for training data goes down with epochs but for validation data it first falls and then goes on increasing. Accuracy plot reflects this. Accuracy for training data finally settles down at .94 but for validation data its around 0.08.
Basically its case of over fitting.
I am using learning rate of 0.005 and dropout of .25.
What measures can I take to get better accuracy for validation? Is it possible that sample size for each class is too small and I may need data augmentation to have more data points?
Hard to say what could be the reason. First you can try classical regularization techniques like reducing the size of your model, adding dropout or l2/l1-regularizers to the layers. But this is more like randomly guessing the models hyperparameters and hoping for the best.
The scientific approach would be to look at the outputs for your model and try to understand why it produces these outputs and obviously checking your pipeline. Did you had a look at the outputs (are they all the same)? Did you preprocess the validation data the same way as the training data? Did you made a stratified train/test-split, i.e. keeping the class distribution the same in both sets? Is the data shuffles when you feed it to your model?
In the end you have about ~85 images per class which is really not a lot, compare CIFAR-10 resp. CIFAR-100 with 6000/600 images per class or ImageNet with 20k classes and 14M images (~500 images per class). So data augmentation could be beneficial as well.
I am working on a project where I use Spark Mllib Linear SVM to classify some data (l2 regularization). I have like 200 positive observation, and 150 (generated) negative observation, each with 744 features, which represent the level of activity of a person in different region of a house.
I have run some tests and the "areaUnderROC" metric was 0.991 and it seems that the model is quite good in classify the data that I provide to it.
I did some research and I found that the linear SVM is good in high dimensional data, but the problem is that I don't understand how something linear can divide my data so well.
I think in 2D, and maybe this is the problem but looking at the bottom image, I am 90% sure that my data looks more like a non linear problem
So it is normal that I have good results on the tests? Am I doing something wrong? Should I change the approach?
I think you question is about 'why linear SVM could classfy my hight Dimensions data well even the data should be non-linear'
some data set look like non-linear in low dimension just like you example image on right, but it is literally hard to say the data set is definitely non-linear in high dimension because a nD non-linear may be linear in (n+1)D space.So i dont know why you are 90% sure your data set is non-linear even it is a high Dimension one.
At the end, I think it is normal that you have a good test result in test samples, because it indicates that your data set just is linear or near linear in high Dimension or it wont work so well.Maybe cross-validation could help you comfirm that your approach is suitable or not.
Would appreciate your input on this. I am constructing a regression model with the help of genetic programming.
If my RMSE on test data is (much) lower than my RMSE on training data for a 1:5 ratio of data, should I be worried?
The test data is drawn randomly without replacement from a set of 24 data points. The model was built using genetic programming technique so the number of features, modeling framework etc vary as I minimize the training RMSE regularized by the number of nodes in the GP tree.
Is the model underfitted? Or should I have minimized MSE instead of RMSE (I thought it would be the same as MSE is positive and the minimum of MSE would coincide with the minimum of RMSE assuming the optimizer is good enough to find the minimum)?
Tks
So your model is trained on 20 out of 24 data points and tested on the 4 remaining data points?
To me it sounds like you need (much) more data, so you can have a larger train and test sets. I'm not surprised by the low performance on your test set as it seems that your model wasn't able to learn from such few data. As a rule of thumb, for machine learning you can never have enough data. Is it a possibility to gather a larger dataset?