A framework that incoporate android camera and ncnn deep learning model deployment in a single app. The default android samples provided from ncnn repo work on a single image. This sample shows how to open android camera and feed frames to ncnn models for inference continuously.
If you want to knoe about ncnn, please refer to: https://github.com/Tencent/ncnn
About the yolo model included with the repo: Model name: yolov2-tiny_voc For more info on yolo, please go to: https://pjreddie.com/darknet/yolov2/ and https://github.com/thtrieu/darkflow I used darknet2caffe to convert yolo to caffe, then used caffe2ncnn to convert the model into NCNN deployable model. Also ncnn2mem is used to fruther convert the model format to binary for use with the android framework: NCNNCam.
The Camera part comes from Google's tensorflow android example and I adapted it for use in this framework. Tensorflow repo is available from: https://github.com/tensorflow/tensorflow
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。