On Wednesday, November 8, 2017 at 2:26:43 AM UTC-5, David Sixela wrote: > > My bad i didn't know you we're working with videos, i thought you had > images like the one you shared. >
A video is just a sequence of images. It seems to me the OpenCV's Scene Text Detector might be useful here. If you wanted to get fancy, you could first identify and track the car to help constrain the area that you were looking at. Once you've identified the region with the 1-3 digit number you'll probably have to convert the contrasting colors to black-and-white if black on white isn't always the case. Tom > Le mercredi 8 novembre 2017 02:54:14 UTC+2, Rod Lisson a écrit : >> >> Thanks David, >> Unfortunately I will be processing these from a video feed and the car >> number could be anywhere in the image. I kind of need to test about 10 >> images as a car goes past to see if I can recognise a number. No human >> interaction possible. >> This may not be the right tool for my job. >> >> Regards, >> Rod >> >> On Wednesday, November 1, 2017 at 11:16:09 PM UTC+8, David Sixela wrote: >> >>> Hi Rod, >>> I think you should do some image processing (by using OpenCv for >>> example) to segment the text area from the rest of the image and then pass >>> the output image result in tesseract. >>> >>> Hope this helps you. >>> >>> Le mercredi 1 novembre 2017 10:06:25 UTC+2, Rod Lisson a écrit : >>>> >>>> HI All, >>>> Just found tesseract, installed and performed OCR on the demo files. >>>> Works well. >>>> Now before I invest too much time, am I asking too much of this >>>> software to recognize numbers on the side of race cars as per the attached >>>> image? >>>> Out of the box it doesn't get any text from the image, but not sure if >>>> training will help or if I just need some parameters. >>>> >>>> Basically I'm trying to get race numbers from hundreds of car images. >>>> The fonts should all be close to this example due to regulations. >>>> >>>> Thanks in advance for any advice. >>>> >>>> -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/a65b5879-acbc-41cd-b8b2-caea039be7f5%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
