I want to build my own Object Detection algorithm using HOG feature. Since OpenCV has its own framework to do this for pedestrian detect, I think I can just modify some of the parameters to customize my own one. But I have a few questions about it after reading this.
1.Prepare my own dataset:
Do I have to make all the pos and neg images the same size? Sometimes resize the image may lead to image deformation and affect the hog result. If not, then I have to change the HOG parameters to adapt every image(for example: set the window size = image size and generate a 3780 vector). which one is better.(I prefer the second one)
2.For training the SVM.
In opencv, I think they use SVMLight for training which has been integrated in OpenCV. Can I use Libsvm or other package (compatible with function hog.setSVMDetector())?
3.For hog.detectMultiScale() function
After doing all these above, can I get the result (rectangles) just calling this function?
Thanks!
Related
My intention is to build a classifier that correctly classify the image ROI with the template that I have manually extracted
Here is what I have done.
My first step is to understand what should be done to achieve the above
I have realized I would need to create the representation vectors(of the template) through research from the net. Hence I have used Bag of words to create the vocabulary
I have used and rewritten the Roy's project to opencv 3.1 and also used his food database. On seeing his database, I have realised that some of the image contain multiple class type. I try to clip the image so that each training image only contains one class of item but the image are now of different size
I have tried to run this code. The result is very disappointing. It always points to one class.
Question I have?
Is my step in processing the training image wrong? I read around and some posts suggest the image size must be constant or at least the aspect ratio. I am confused by this. Is there some tools available for resizing samples?
It does not matter what is the size of the sample images, since Roy's algorithm uses local descriptors extracte from nearby points of interest.
SVM is linear regression classifier and you need to train different SVM-s for each class. For each class it will say whether it's of that class or the rest. The so called one vs. rest.
I want to extract HOG features using the OpenCV Module. I just want the features, not the whole svm detection process.
To speed up things, I would like to use the HOGCache object.
My question: is this possible to use the HOGCache on a specified zone of the image, and then call the HOG compute method (or a changed version, I guess) to extract the HOG vector at a specified location ?
The HOG compute has to be called at several locations, and not using the cache gives me a really slow process.
I guess this is possible since the svm detector use the cache, but with a a sliding window.
Thx for your help
I have written an object classification program using BoW clustering and SVM classification algorithms. The program runs successfully. Now that I can classify the objects, I want to track them in real time by drawing a bounding rectangle/circle around them. I have researched and came with the following ideas.
1) Use homography by using the train set images from the train data directory. But the problem with this approach is, the train image should be exactly same as the test image. Since I'm not detecting specific objects, the test images are closely related to the train images but not essentially an exact match. In homography we find a known object in a test scene. Please correct me if I am wrong about homography.
2) Use feature tracking. Im planning to extract the features computed by SIFT in the test images which are similar to the train images and then track them by drawing a bounding rectangle/circle. But the issue here is how do I know which features are from the object and which features are from the environment? Is there any member function in SVM class which can return the key points or region of interest used to classify the object?
Thank you
I have been tasked to use OpenCV and C++
Read a set of videos for creating a set of images/learning.
Classify objects seen in the videos
Label the images
test against series of test videos to check objects were identified as expected. draw a rectangle around them and label.
I am new to OpenCV however happy to program in C++ as soon as approach is formed. I am also planning to write my own functions at a later stage.
I need your help in formning right way of solution approach as I have to identify household objects [cup, soft toy, phone, camera, keyboard) from a stream of video and then test on another stream of video. The original video has depth information as well but not sure how to use it to my benefit.
Read about Support vector machine (SVM) , Feature extraction (e.g. SIFT/SURF) , SVM training and SVM testing. And, for drawing Rectangle, read about findContour(), drawContour() in openCV.
Approach:
Detect objects (e.g. car/plane etc.). Store the points of its contours
Extract some features of that object using SIFT/SURF
Based upon the extracted features, classify the object using SVM (the input for SVM will be the extracted features)
And if the SVM says -Yes! it is a car. Then, draw a rectangle around it using the points of its contour which you had stored in first step.
I'm trying to build a lightweight object recognition system using ORB for feature extraction and LDA for classification. But I'm running into an issue do to the varying size of extracted features.
These are my steps:
Extract keypoints using ORB.
Extract trainable features in the image by grouping the keypoints.
(example of whats being extracted: http://imgur.com/gaQWk)
Train the recognizer with the extracted features. (This is where problems arise)
Classify objects in an image from the wild.
If I attempt to create a generalized matrix using cv::gemm, I get an exception due to the varying sizes. My first thought was to just to normalize all the images by resizing them, but this causes a lot of accuracy issues when objects have similar small features.
Is there any solution to this? Is LDA an appropriate method for this? I know it's commonly used with facial recognition algorithms such as fisherfaces.
LDA requires fixed length features, as do most optimization and machine learning methods. You could resize the image patches to be a fixed size, but that is probably not going to be a good feature. Normally people use a scale invariant feature such as SIFT. You also might try a color histogram, or some variation of edge detection and spatial histogram binning such as a GIST vector.
It's hard to say if LDA is an appropriate method for this without knowing what you hope to accomplish. You might also look into using SVM, some form of boosting, or just plain nearest neighbor with a large training set.