Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 8 years ago.
Improve this question
Now I'm doing human face gender recognition using javacv, The correctness is not very satisfactory for only about 70%. The training set I'm using is some asia human face images I get from internet, and I both used geometry info based training & classification and appearance based info.
The geometry info is produced by choosing some face feature points ratios such like faceheight/facewidth etc and trained by a fisher recognizor, while the appearance info is trained by a LBP face recognizor.
My question is: what is the key point to improve the correctness of gender recognition? Can someone share some experiences?
Thanks~
Though I don't consider myself to be the expert in CV, but I have dabbled a bit to point you to the right literature.
Look at this paper for a survey of work used in human gender recognition from face images:
http://arxiv.org/pdf/1204.1611.pdf
Also look at following papers:
http://www.cse.unr.edu/~bebis/GenderRecognitionIWSSIP12.pdf
http://tdlc.ucsd.edu/research/publications/Nestor_Tarr_Gender_Recognition.pdf
http://www.ijarcce.com/upload/2013/june/43-Hadeel%20Alrashed%20-facial%20gender%20recognition%20using%20eyes%20images.pdf
Related
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 2 years ago.
Improve this question
Can someone please give me some suggestions on which feature selection techniques for gene classification should I use?
The major problem to work with gene expression data, with a large number of dimensions and small sample size. Instead of standard feature extraction/selection algorithms, generally, kernel-based feature selection algorithms are applied to gene expression data such as KBMTL(kernelized Bayesian multitask learning), NDR(nonlinear dimensionality reduction) or regularized linear methods such as LASSO and Elastic-net.
You can check these papers to learn more about how to make efficient feature selection on gene expression data.
paper1
paper2
paper3
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 2 years ago.
Improve this question
I am working on building a custom facial recognition for our office.
I am planning to use Google FaceNet,
Now my question is that you can find or create your own version of facenet model in keras or pytorch there's no issue in that, but regarding creating dataset ,I want to know what are the best practices to capture photo of person when I don't have any prior photo of that person,all I have is a camera and a person ,should I create variance in by changing lightning condition or orientation or face size ?
A properly trained FaceNet model should already be somewhat invariant to lighting conditions, pose and other features that should not be a part of identifying a face. At least that is what is claimed in a draft of the FaceNet paper. If you only intend to compare feature vectors generated from the network, and intend to recognize a small group of people, your own dataset likely does not have to be particulary large.
Personally I have done something quite similar to what you are trying to achieve for a group of around ~100 people. The dataset consisted of 1 image per person and I used a 1-N-N classifier to classify the generated feature vectors. While I do not remember the exact results, it did work quite well. The pretrained network's architecture was different from FaceNet's but the overall idea was the same though.
The only way to truly answer your question though would be to experiment and see how well things work out in practice.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 4 years ago.
Improve this question
Hello everyone I'm building a license plate detection model in Tensorflow. I built a function that chooses a license plate at random from a collection of ~5000 plates and puts it in a random place in on a random background and saves the coordinates. At first I thought to generate about 40K images this way and train the network on with the generated data. But wouldn't it be a good idea to just continiously keep generating new data to feed to the network and basically eliminate any chance of it getting overfitted?
This is an excellent way to train it on how to spot the discontinuities around a superimposed yellow / white / blue rectangle, but maybe not such a great way of teaching it to spot a real license plate. If you've got a good way of procedurally generating images then great! but be warned.
It might spot the wrong pattern.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 8 years ago.
Improve this question
Hi i want to predict health level(High,medium,low) in leaf using image processing and data mining.So far i thought using extract colors from leaf using Bayes algorithm to predict healthy of leaf. and data mining part have completed now.but i need extra features for prediction.we only used orchid leaf.So i can't use vain structure.Can anyone help me to what are the other features can be extracted from leaf for identify health level of leaf.Any idea or comments help me to improve my project. Thanks
There are many possible approaches to a problem like this. One common method is the bag-of-features model. Take a look at this example using the Computer Vision System Toolbox in MATLAB.
Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 8 years ago.
Improve this question
I wanna experiment k-means clustering method on different kind of images, so I am trying to find different kind of images used in image segmentation such as MRI images.I want to gather some more categories.
Any suggestion would be gratefully appreciated.
Although this is not the correct place for asking your question, to help you ,Image segmentation has a wide range of application including segmenting Satellite imagery
and Medical Imaging images, Texture Recognition, Facial Recognition System, Automatic Number Plate Recognition, and a lot of other machine vision applications.