Download full-text PDF |
Source |
|---|
Publication Analysis
Top Keywords
Similar Publications
Sparse-Coding Variational Autoencoders.
Neural Comput
November 2024
Princeton Neuroscience Institute, Princeton University, Princeton, NJ 08544, U.S.A.
Article Synopsis
- - The sparse coding model suggests that our visual system uses a limited number of features to efficiently process complex natural images, but it faced issues with complicated computation and uncertainty in fitting.
- - A new approach called the sparse coding variational autoencoder (SVAE) combines the sparse coding model with a deep neural network for better recognition and fit to data by maximizing the evidence lower bound (ELBO).
- - The SVAE differs from traditional variational autoencoders by having an overcomplete latent representation, a sparse prior instead of a Gaussian one, and a simpler linear decoder, and it shows improved performance on natural image data while capturing crucial neuron properties in early visual processing.
Event-enhanced Snapshot Mosaic Hyperspectral Frame Deblurring.
IEEE Trans Pattern Anal Mach Intell
September 2024
Article Synopsis
- Snapshot Mosaic Hyperspectral Cameras (SMHCs) face challenges in motion perception due to narrow-band spectral filters, which can cause blurry images.
- The paper proposes a hardware-software collaboration that combines SMHCs with neuromorphic event cameras to enhance frame clarity and recover spectral details using a new model called spectral-aware Event-based Double Integral (sEDI).
- A Noise Awareness training framework, along with a specialized Event-enhanced Hyperspectral frame Deblurring Network (EvHDNet), improves the robustness and effectiveness of image deblurring, showing better results than current top methods.
Mine Your Own Anatomy: Revisiting Medical Image Segmentation With Extremely Limited Labels.
IEEE Trans Pattern Anal Mach Intell
September 2024
Recent studies on contrastive learning have achieved remarkable performance solely by leveraging few labels in the context of medical image segmentation. Existing methods mainly focus on instance discrimination and invariant mapping (i.e.
View Article and Find Full Text PDFExamining the challenges of blood pressure estimation via photoplethysmogram.
Sci Rep
August 2024
Microsoft Research, Redmond, USA.
The use of observed wearable sensor data (e.g., photoplethysmograms [PPG]) to infer health measures (e.
View Article and Find Full Text PDF: a -based indexing algorithm for kilohertz serial crystallography.
J Appl Crystallogr
August 2024
Swiss Data Science Center, Paul Scherrer Institute, Villigen, Switzerland.
Serial crystallography (SX) involves combining observations from a very large number of diffraction patterns coming from crystals in random orientations. To compile a complete data set, these patterns must be indexed ( their orientation determined), integrated and merged. Introduced here is (-powered robust optimization) , a robust and adaptable indexing algorithm developed using the framework.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!