Italy


Situation in Image compression:  Although existing methods based on a fully convolutional network have greatly improved the performance, the reconstructed results still do not show satisfactory on the structural similarity and statistical distribution similarity of the whole image.
CNN&#39;s Shortcoming: Overdownsampling thus losing details
Proposal: In this work, we propose an GAN based dense Auto-encoder (GDAE) and Semantic-perceptual residual compensation method (SPRC) to improve the coarse results generated by front CNN and compensate the details of reconstructed image.

ICPR 2020

Adaptive Image Compression Method Using GAN based Semantic-Perceptual Residual Compensation

semantic perception

image compression

deep learning

Promoted by the Brazilian Computer Society (SBC), the Brazilian Symposium on Multimedia and Web (WebMedia) is the main event of the theme in Brazil and an excellent opportunity for scientific and technical exchange among students, researchers and professionals in the areas of Multimedia, Hypermedia and Web. This year we will be celebrating to 26th edition of WebMedia, which will be held online from november 30th to December 4th .

Webmedia 2020’s programming includes:

– Technical sessions for the presentation of scientific papers
– National and international keynotes
– Panel with companies
– Workshops
– Short Courses
– Open community meetings

The topics of interest for submitting papers are:

– Web and Social Networks
– Multimedia
– Digital TV, Ubiquitous and Mobile Computing

Besides the main track, WebMedia 2020 will have in parallel the Tools and Applications Workshop, works Tools and Applications Workshop, the Workshop on Undergraduate Research Work and the Workshop The Future of Interactive.

The WebMedia 2020 organization is under the responsibility of the Programa de Pós-graduação Doutorado em Ciência da Computação – Associação UFMA/UFPI (DCCMAPI), do Programa de Pós-graduação em Ciência da Computação/ CCET (PPGCC), do Laboratory of Advanced Web Systems/UFMA (LAWS) e do Laboratório de Sistemas Multimídia/UFMA (TELEMÍDIA).).

#PartiuWebMedia2020 #WebMedia2020

WebMedia 2020

Promoted by the Brazilian Computer Society (SBC), the Brazilian Symposium on Multimedia and Web (WebMedia) is the main event of the theme in Brazil and an excellent opportunity for scientific and technical exchange among students, researchers and professionals in the areas of Multimedia, Hypermedia and Web.


Situation in Image compression:  Although existing methods based on a fully convolutional network have greatly improved the performance, the reconstructed results still do not show satisfactory on the structural similarity and statistical distribution similarity of the whole image.
CNN's Shortcoming: Overdownsampling thus losing details
Proposal: In this work, we propose an GAN based dense Auto-encoder (GDAE) and Semantic-perceptual residual compensation method (SPRC) to improve the coarse results generated by front CNN and compensate the details of reconstructed image.

poster

ICPR 2020 is the premier international conference on Pattern Recognition, gathering leading researchers from around the world to present their latest achievements in Pattern Analysis and Recognition, applied Artificial Intelligence and Machine Learning. The topic of the year is " Putting Artificial intelligence to work on Patterns". 
For the first time, ICPR has employed a two-round review process similar to journal submissions, with new papers allowed to be submitted in either the first or second round and papers submitted in the first round and not accepted allowed to be revised and re-submitted for second round review. 

In this edition of the **ICPR 2020** Conference, there are 1407 Oral/Poster presentations organized in five Tracks: 
- Artificial Intelligence, Machine Learning for Pattern Analysis; 
- Biometrics, Human Analysis and Behavior Understanding; Computer Vision, Robotics and Intelligent Systems; 
- Document and Media Analysis; and 
- Image and Signal Processing. 

The **ICPR 2020** Main Conference program also includes 7 Keynotes from prominent world-renowned scientists, 8 Tutorials, 28 Demos, 8 Challenges, 2 Panels, as well as an Industry/Sponsor Exhibition. In addition, 41 Workshops on a range of important current and emerging research topics complete the ICPR 2020 program. 


In this paper we propose a novel method for location recognition and prediction in smart homes based on semi-supervised learning. We use data collected from low-resolution depth video cameras installed in four apartments with older adults over 70 years of age, and collected during a period of one to seven weeks. The location of the person in the depth images is detected by a person detection algorithm adapted from YOLO (You Only Look Once). The locations extracted from the videos are then clustered using K-means clustering. Sequence prediction algorithms are used to predict the next cluster (location) based on the previous clusters (locations). The accuracy of predicting the next location is up to 91\%, a significant improvement compared to the case where binary sensors are placed in the apartment based on human intuition. The paper presents an analysis on the effect of the memory length (i.e. the number of previous clusters used to predict the next one), and on the amount of recorded data required to converge.

Location Prediction in Real Homes of Older Adults based on K-Means in Low-Resolution Depth Videos

This paper focuses on the problem of visual saliency prediction, predicting regions of an image that tend to attract human visual attention, under a constrained computational budget. We modify and test various recent efficient convolutional neural network architectures like EfficientNet and MobileNetV2 and compare them with existing state-of-the-art saliency models such as SalGAN and DeepGaze II both in terms of standard accuracy metrics like Area Under Curve (AUC) and Normalized Scanpath Saliency (NSS), and in terms of the computational complexity and model size. We find that MobileNetV2 makes an excellent backbone for a visual saliency model and can be effective even without a complex decoder. We also show that knowledge transfer from a more computationally expensive model like DeepGaze II can be achieved via pseudo-labelling an unlabelled dataset, and that this approach gives result on-par with many state-of-the-art algorithms with a fraction of the computational cost and model size.

FastSal: a Computationally Efficient Network for Visual Saliency Prediction


In this paper we define a deep multi-stage architec-
ture for automated landmarking of craniomaxillofacial (CMF)
CT images. Our model is composed of three subnetworks that
first localize, on reduced-resolution images, areas where land-
marks may be found and then refine the search, at full-resolution
scale, through a hierarchical structure aiming at increasing the
granularity of the investigated region. The multi-stage pipeline is
designed to deal with full resolution data and does not require any
additional pre-processing step to reduce search space, as opposed
to existing methods that can be only adopted for searching
landmarks located in well-defined anatomical structures (e.g.,
mandibles). The automated landmarking system is tested on
identifying landmarks located in several CMF regions, achieving
an average error of 0.8 mm, significantly lower than expert
readings. The proposed model also outperforms baselines and
is on par with existing models that employ additional upstream
segmentation, on state-of-the-art benchmarks.


Deep Recurrent-convolutional Model for Automated Segmentation of Craniomaxillofacial Ct Scans 


Malware evolves rapidly over time, which makes existing solutions being ineffective in detecting newly released malware. Machine learning models that can learn to capture malicious patterns directly from the data play an increasingly important role in malware analysis. However, traditional machine learning models heavily depend on feature engineering. The extracted static features are vulnerable as hackers could create new malware with different feature values to deceive the machine learning models. In this paper, we propose an end-to-end malware detection framework consisting of convolutional neural network, autoencoder and neural decision trees. It learns the features from multiple domains for malware detection without feature engineering. In addition, since anti-virus products should have a very low false alarm rate to avoid annoying users, we propose a special loss function, which optimizes the recall for a fixed low false positive rate (e.g., less than 0.1%). Experiments show that the proposed framework has achieved a better recall than the baseline models, and the derived loss function also makes a difference.

Malware Detection by Exploiting Deep Learning over Binary Programs

Brief introduction about the motivation, method, experiments of our work

Learning a Dynamic High-Resolution Network for Multi-Scale Pedestrian Detection


In recent years, skin cancer has become one of the most common cancers. Among all types of skin cancers, melanoma is the most fatal one and many people die of this disease every year. Early detection can greatly reduce the death rate and save more lives. Skin lesions are one of the early symptoms of melanoma and other types of skin cancer. So accurately recognizing various skin lesions in early stage is of great significance. There have been lots of existing works based on convolutional neural networks (CNN) to solve skin lesion classification but seldom do they involve the similarity among different lesions. For example, we find that some lesions like melanoma and nevi look similar in appearance which is hard for neural network to distinguish categories of skin lesions. Inspired by fine-grained image classification, we propose a novel network to distinguish each category accurately. In our paper, we design an effective module, distinct region proposal module (DRPM), to extract the distinct regions from each image. Spatial attention and channel-wise attention are both utilized to enrich feature maps and guide the network to focus on the highlighted areas in a weakly-supervised way. In addition, two preprocessing steps are added to ensure the network to get better results. We demonstrate the potential of the proposed method on ISIC 2017 dataset. Experiments show that our approach is effective and efficient.

Skin Lesion Classification Using Weakly-supervised Fine-grained Method


Most of the state-of-the-art automatic music transcription (AMT) models break down the main transcription task into sub-tasks such as onset prediction and offset prediction and train them with onset and offset labels. These predictions are then concatenated together and used as the input to train another model with the pitch labels to obtain the final transcription. We attempt to use only the pitch labels (together with spectrogram reconstruction loss) and explore how far this model can go without introducing supervised sub-tasks. In this paper, we do not aim at achieving state-of-the-art transcription accuracy, instead, we explore the effect that spectrogram reconstruction has on our AMT model. Our proposed model consists of two U-nets: the first U-net transcribes the spectrogram into a posteriorgram, and a second U-net transforms the posteriorgram back into a spectrogram. A reconstruction loss is applied between the original spectrogram and the reconstructed spectrogram to constrain the second U-net to focus only on reconstruction. We train our model on three different datasets: MAPS, MAESTRO, and MusicNet. Our experiments show that adding the reconstruction loss can generally improve the note-level transcription accuracy when compared to the same model without the reconstruction part. Moreover, it can also boost the frame-level precision to be higher than the state-of-the-art models. The feature maps learned by our U-net contain gridlike structures (not present in the baseline model) which implies that with the presence of the reconstruction loss, the model is probably trying to count along both the time and frequency axis, resulting in a higher note-level transcription accuracy.

The Effect of Spectrogram Reconstruction on Automatic Music Transcription: An Alternative Approach to Improve Transcription Accuracy

Identifying vehicles across a network of cameras with non-overlapping fields of view remains a challenging research problem due to scene occlusions, significant inter-class similarity and intra-class variability. In this work, we propose an end-to-end multi-level re-identification network that is capable of successfully projecting same identity vehicles closer to one another in the embedding space, compared to vehicles of different identities. Robust feature representations are obtained by combining features at multiple levels of the network. As for the learning process, we employ a recent state-of-the-art structured metric learning loss function previously applied to other retrieval problems and adjust it to the vehicle re-identification task. Furthermore, we explore the cases of image-to-image, image-to video and video-to-video similarity metric. Finally, we evaluate our system and achieve great performance on two large-scale publicly available datasets. Compared to most existing state-of-art approaches, our approach is simpler and more straightforward, utilizing only identity-level annotations, while avoiding post-processing the ranking results (re-ranking) at the testing phase.

Multi-level Deep Learning Vehicle Re-identification using Ranked-based Loss Functions

In this work, loop-closure detection from LiDAR scans is defined as an image re-identification problem. Re-identification is performed by computing Euclidean distances of a query scan to a gallery set of previous scans. The distances are computed in a feature embedding space where the scans are mapped by a convolutional neural network (CNN). The network is trained using the triplet loss training strategy. In our experiments we compare different backbone networks,
variants of the triplet loss and generic and LiDAR specific data augmentation techniques. With a realistic indoor dataset the best architecture obtains the mean average precision (mAP) above 0.94.

Loop-closure detection by LiDAR scan re-identification

Stochastic Label Refinery is a simple method to generate better soft label and improve the model performance.

Downloads

Next from ICPR 2020

Location Prediction in Real Homes of Older Adults based on K-Means in Low-Resolution Depth Videos

Similar lecture

Decoder-Side Quality Enhancement for JPEG Images Using a Deep Learning-Based Prediction Model for Quantized DCT Coefficients