Computer-Vision | Computer Vision - Impemented algorithms | Machine Learning library

I'm thinking about coding a computer-vision program that can take an image as input and identify whether a certain symbol exists inside that image. For example, the input could be a guy holding a paper with International Symbol of Access printed, and the code would process this image and correctly identify the presence of this symbol.

I'm kind of new to this field, and done some research before asking the question. From my understanding I should train a model on a dataset that consists of images which do and do not contain the said symbol. The inputs to the model should be the symbol that I'm trying to detect coupled with the images from the datasets, and the output during the training will be either 1 or 0 depending on the image. The problem that I'm facing is, I don't know whether or not my choice of input/outputs is correct, and also I don't know how to generate my dataset. My project will be using a custom symbol to be detected in the images. I've tried looking up similar researches in the internet, but haven't been able to find any. I plan using tensorflow since I've worked with it earlier.

If you've come across similar projects, I'd really like to hear some guidance or even references that I can work with. Thanks in advance.

I am using Microsoft Computer Vision API for OCR processing and I noticed that they are getting charged as S3 transactions instead of S2 in my bill.

I'm using the .NET SDK and the API I am using is this one. https://docs.microsoft.com/en-us/dotnet/api/microsoft.azure.cognitiveservices.vision.computervision.computervisionclientextensions.readasync?view=azure-dotnet

I have also confirmed that the actual REST API the SDK calls is the following POST /vision/v3.2/read/analyze https://centraluseuap.dev.cognitive.microsoft.com/docs/services/computer-vision-v3-2/operations/5d986960601faab4bf452005

According to documentation, that should be the OCR Read API, am I correct? https://docs.microsoft.com/en-us/azure/cognitive-services/computer-vision/vision-api-how-to-topics/call-read-api

I am puzzled as to why my calls are getting charged as S3 instead of S2. This is important for me because S3 is 50% more expensive than S2. Using the Pricing Calculator, 1000 S2 transactions is $1, whereas 1000 S3 transactions is $1.5. https://azure.microsoft.com/en-us/pricing/calculator/?service=cognitive-services

What's the difference between OCR and "Describe and Recognize Text" anyways? OCR (Optical Character Recognition) by definition must recognize text. I am calling the Read API without any of the optional parameters so I did not ask for "Describe" hence the call should be S2 feature rather than S3 feature I think.

I already posted this question at Microsoft Q&A but I thought SO might get more traffic hence help me get an answer faster. https://docs.microsoft.com/en-us/answers/questions/689767/computer-vision-api-charged-as-s3-transaction-inst.html

I'm trying to implement a neural network to generate sentences (image captions), and I'm using Pytorch's LSTM (nn.LSTM) for that.

The input I want to feed in the training is from size batch_size * seq_size * embedding_size, such that seq_size is the maximal size of a sentence. For example - 64*30*512.

After the LSTM there is one FC layer (nn.Linear). As far as I understand, this type of networks work with hidden state (h,c in this case), and predict the next word each time.

My question is- in the training - do we have to manually feed the sentence word by word to the LSTM in the forward function, or the LSTM knows how to do it itself?

My forward function looks like this:

So today I updated Android Studio to:

I'm using Huawei image segmentation for background removal from images. This code work perfectly fine on debug build but it does not work on a release build. I don't understand what could be the case.

Code:

I have scanned documents which are printed by different inkjet-printers (Epson, HP, Canon and so on). Each photo has a very high quality (like 1,6GB) and you can zoom in and see the halftone of the picture which is using a frequency modulation.

My task is to do a feature extraction based on the grid dots, patterns of the grid, distance of the dots etc.

The relevant features are the size of these dots (each printers print different size of these dots - have to calculate the mean and standard deviation).

Later I will have to train a model with ML and the trained model should classify a print to a specific printer (so basically this print belongs to printer XYZ).

But for now I am struggling with the feature engineering already and the pre processing stuffs as this is my first computer-vision project actually and I am not so familar with opencv.

I have an idea and my plan is to a binary transformation with opencv of the images to determine the edges (Edge detection) via Sobel or Prewitt filter or whatsoever. So I think I have to put some blur and then a edge detection maybe?

I am not sure if this is the right approach, so that's why I ask here, what do you think? I would be happy if you can give me some hints or steps for the best or a good approach.

I am building a CNN model using Resnet50 to identify to classify 5 objects. The images of the objects were taken on my desk so there is a portion of my desk in every object. The code to initialize the model is, like so,

The app was working perfectly with the previous version :

I get this error when I try to run a code utilizing Azure Computer Vision.