This is Faisal Alamri, I am a computer science student. I am working on matlab, particularly in object detection.
My project is about detecting multiple object classes (object classes detection), where the dataset that I am using is MSCOCO.
I am looking for help to do this detection, I have tried matlab Faster RCNN detection method, but I am still facing some errors.
I need the following
1/ get faster rcnn trained on coco dataset,
Or 2/ if there is an already trained model (pre-trained model) that I can use it.
Accuracy should be high,
I need to have the detection results trained on coco dataset.
Please note the following:
1/ annotation of coco dataset is originally in .json file, which can be downloaded from the link below
2/ I have converted the annotation from .json to .xml files following the format of PASCAL dataset.
3/ I have access to two servers, where GPUs for both are 1) 2x NVIDIA Tesla K80, and server 2) 2x NVIDIA Tesla K40, they can be used anytime for training. However, if you want to do the training on your side, it is totally up to you.
Please also note that: I prefer to work only with matlab, but if it cheaper or more efficient to have the training done on any different programming language (e.g., python) I don’t mind, where I need the results to be accessible by matlab, saved in .mat file.
I am using coco2014, to download dataset, please go to ([login to view URL] )
For further information about Faster RCNN, please have a look at ([login to view URL])
Thank you very much for your help and time,
I am also looking for someone who can do the above described job, plus be with me when needed, because I have just started this project, and I will need further help
The cheaper, the better: I don’t have any budget in mind, but I would say, I am a student, and you can imagine how hard it is to pay for my dairy expenses, so again the cheaper the better.
Thank you again,
13 frilansere byr i snitt £396 for denne jobben
Hi I would love to do your task. If you hire me, i won't let you down. I can provide you all these things with unlimited revisions till the satisfactorily completion of [login to view URL] for your message .Thank you