Wei Liu, Dragomir Anguelov, Dumitru Erhan
Christian Szegedy, Scott Reed, Cheng-Yang Fu, Alexander C. Berg
2015 · arXiv
We present a method for detecting objects in images using a single deep neural network. Our approach, named SSD, discretizes the output space of bounding boxes into a set of default boxes over different aspect ratios and scales per feature map location. At prediction time, the network generates scores for the presence of each object category in each default box and produces adjustments to the box to better match the object shape. Additionally, the network combines predictions from multiple feature maps with different resolutions to naturally handle objects of various sizes. Our SSD model is simple relative to methods that require object proposals because it completely eliminates proposal generation and subsequent pixel or feature resampling stage and encapsulates all computation in a single network. This makes SSD easy to train and straightforward to integrate into systems that require a detection component. Experimental results on the PASCAL VOC, MS COCO, and ILSVRC datasets confirm that SSD has comparable accuracy to methods that utilize an additional object proposal step and is much faster, while providing a unified framework for both training and inference. Compared to other single stage methods, SSD has much better accuracy, even with a smaller input image size. For $300\times 300$ input, SSD achieves 72.1% mAP on VOC2007 test at 58 FPS on a Nvidia Titan X and for $500\times 500$ input, SSD achieves 75.1% mAP, outperforming a comparable state of the art Faster R-CNN model. Code is available at https://github.com/weiliu89/caffe/tree/ssd .
Metadata: arXiv. Abstracts retain their original authors’ rights.
Original HN comments, with labels from simple text rules. A label describes the whole comment and may miss context or irony. Read the discussion to make up your own mind. About the labels
3 selected excerpts · 3 total mentions
I don't think pjreddie invented SSD, at least he's not on the author list. https://arxiv.org/abs/1512.02325 Yolo is great though and a fork is very actively maintained on GitHub at alexeyAB/darknet. Darknet itself is a nice library that doesn't need any dependencies aside from CUDA. It's all C (albeit uncommented) and quite easy to walk through if you want to see at a low level how deep learning frameworks work.
For detecting objects in video, take a look at: https://github.com/tensorflow/models/tree/master/research/ls... It uses SSD object detector (https://arxiv.org/abs/1512.02325 , https://lambdalabs.com/blog/how-to-implement-ssd-object-dete...) on each frame, then runs LSTM on top of it to collect information across video frames. The trained model can be run on mobile devices via Tflite.
There are quite a few papers that discuss the application R-CNN (Region-based Convolutional Neural Network) for object detection and classification [1-3]. On websites such as Medium, you can find lots of blog posts on object detection using popular frameworks such as Tensorflow (e.g. https://towardsdatascience.com/how-to-train-your-own-object-...). [1] Faster R-CNN: https://arxiv.org/abs/1506.01497 [2] Single Shot Multibox: https://arxiv.org/abs/1512.02325 [3] YOLO9000: https://arxiv.org/abs/1612.08242
More from the same topics.
Tom B. Brown, Benjamin Mann, Nick Ryder +28
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task.
Chiyuan Zhang, Samy Bengio, Moritz Hardt +2
Despite their massive size, successful deep artificial neural networks can exhibit a remarkably small difference between training and test performance.
Mariusz Bojarski, Davide Del Testa, Daniel Dworakowski +10
We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands.
David Manheim, Scott Garrabrant
There are several distinct failure modes for overoptimization of systems on the basis of metrics.