The main contribution of this work is the proposal of a densely connected convolutional network for semantic segmentation, which strengthens utilization of features and improves segmentation results even with limited training samples. To achieve this, we combine the U-Net network and our resulting system is called Dense-U-Net. Compared to traditional convolutional networks such as U-Net, there are additional concatenation layers between each pair of convolutional layers which have the same size of outputs in our Dense-U-Net, each layer can get the feature-maps of all its preceding layers as inputs while its feature-maps can be passed to all subsequent layers, and a higher segmentation quality can be achieved without a need for increasing the volume of datasets finally. We evaluate our proposed architecture by segmentation accuracy, foreground-restricted rand scoring after border thinning VRand and foreground-restricted information theoretic scoring after border thinning VInfo at the same time, and the results are shown on three different segmentation tasks: ISBI challenge 2012 for segmentation of neuronal structures in electron microscopic stacks, ISBI cell tracking challenge 2014(Glioblastoma-astrocytoma U373 cells) and 2015(HeLa cells), our Dense-U-Net achieves better results than U-Net and several other state-of-the-art networks on all tasks.
Tracking methods based on Correlation Filter have been constantly improved in tracking accuracy and robustness. However, it still challenged in background clutter, rotation changes and occlusion, the drift of the model was one of the main reasons. In this paper, we propose an online sample training method based on Gaussian Mixture Model. The maximum response value, obtained from the convolution of samples and filters, is used to judge the availability of the online samples, which is able to reduce the interference of wrong online samples. Then, through Gaussian Mixture Model, samples are classified to strengthen the diversity of the sample set, which can avoid model drift effectively. Besides, we also propose a model update criterion to enhance the stability of the tracker, and heighten the efficiency of calculation. This criterion is determined by changes of target in scale and displacement. We perform comprehensive experiments on three benchmarks: OTB100, VOT2016 and VOT-TIR2016. Comparing with other trackers, our tracker has better robustness in the condition of background clutter, rotation change and occlusion. Moreover, its speed also maintains real-time performance.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.