ANPR Engine Required

I am excellent at the following : - python - Linux - video recognition (ANPR OR tensor flow) I read your requirements and I am able to deliver great results for your target. if you choose me, I think you won't regret. best regards,

Thanks for your posting I am an image processing expert with full experiences using machine learning such as Yolo, tensorflow, caffe, pytorch, keras and so on... I have developed a lot of object detection and recognition projects such as Pose detection, ANPR, Breast Cancer Detection, Solar panel detection from satellite imagery, Good split and recognition from Ads, Soccer player tracking, Image style transfer and etc... with Java, Android, c#, C++, python. Also, I am very familiar with python and C++ If you response my bid I will send you my image processing Demos and I can finish your task and deliver you perfectly. Also, you can know my ability enough. Please give me your detail Looking forward to hear from you soon. Best Regard.

Hi Pankaj S. Hope you are going well! Thanks for sharing your project requirement with us. I would like to work with you on your project. I first promise to help you with sincere efforts and an honest heart. I have over 7 years of experience in Machine Learning and Artificial Intelligence by using java, python. I have implemented Machine learning and Artificial intelligence in Mobile app and Python. My main skill is: -C++, C#, C, Python/Django, Convolution Neural Network, RNN, Jupyter Notebook Integration, FireWorks, Google Colab, Spike NN, Facebook, etc. -OCR, Web, desktop automation, Image processing, Image recognition, Open CV, Invoice. My goal is a giving the satisfactory to every client. I will provide best communication and high-quality work. I would love to help you and be the person who adapts to what you need. kindly chat me up and give me a fair chance to deliver the best result to you! Best Regards.

Have made ANPR for industrial partner in Mumbai. Check profile for robust real life ANPR solutions. We have achieved success on following areas: 1- Increased FPS for detection by using tflite. 2- ANPR accuracy depends on many factors , its not just your algorithm. * We can discuss all this. 3- Object detection for multiple classes (Car, Bus, Truck, Motorcycle) Have run program on CPU as well as GPU. basically ANPR works best every 5th frame for normal situations. HRSP number plates give best results, however we made it for Indian roads where you see lot of hand painted number plates too. Trained our own model of OCR tesseract for detection , as standard tesseract follows many signs, special symbols. Devised our own "FUZZY LOGIC" to screen vague inputs in ANPR. However, we still believe that process can be improved. Please find details on my profile. Let me know , if this makes sense.

Hi I recently worked on this type of project. Working on embedded device is a challenging task. Raspberry pi would not be a good choice but nano can do. For recognition, I think are some rcnn trained models for license plates available which are faster. Also, do you have datasets for detection and recognition? And which country license plate are you targeting? Anyways let’s talk more via chat and we can discuss. Thank you

Hi I have experience to build this task with grate accuracy so let’s connect to build the entire system Thanks

Hi, Our past experience makes us a good fit for all the points that you have mentioned. Going over one by one with above mentioned points and matching it with our past experience should convince you that we are right for the project. 1. 3D pose estimation had the ability to support multiple cameras (32 maximum limit). it leveraged the information from those cameras to improve its accuracy. 2. Gridge app ran multiple AI models for recognition and was able to process in real time on a mobile phone without requiring any dedicated GPU or internet connectivity. We custom designed all the techniques to make them efficent enough to run on a low end smartphone. 3. One of our unlisted project required us to have high precision and hgh recall (sensitivity) on its OCR. We designed and trained our custom OCR based on LSTM combine with computer vision. 4. Gridge app also required us to classify each object in court and this needed to be donw in real-time and on smartphone. 5. Gridge app was depolyed locally on a smartphone and didn't expect any dedicated GPU support, so we can leverage the same for low powered computing units. 6. 3D pose estimation was designed for sports entertainment which automatically put a requirment of getting high speed movemnets accuratly. We were able to come up with a clever algorithm that could help the AI capture detailin these fast moving/blur images. We would love to talk further with you. Thanks

Even though am in research field, am expertize in solving real world problem, till now i have guided many students in machine learning area and i came to know that i have capability to do data analysis in different ways.