While deep learning strategies achieve outstanding results in computer vision tasks, one issue remains: The current strategies rely heavily on a huge amount of labeled data. In many real-world problems, it is not feasible to create such an amount of labeled training data. Therefore, it is common to incorporate unlabeled data into the training process to reach equal results with fewer labels. Due to a lot of concurrent research, it is difficult to keep track of recent developments. In this survey, we provide an overview of often used ideas and methods in image classification with fewer labels. We compare 34 methods in detail based on their performance and their commonly used ideas rather than a fine-grained taxonomy. In our analysis, we identify three major trends that lead to future research opportunities. 1. State-of-the-art methods are scalable to real-world applications in theory but issues like class imbalance, robustness, or fuzzy labels are not considered. 2. The degree of supervision which is needed to achieve comparable results to the usage of all labels is decreasing and therefore methods need to be extended to settings with a variable number of classes. 3. All methods share some common ideas but we identify clusters of methods that do not share many ideas. We show that combining ideas from different clusters can lead to better performance.

/pdf/a-survey-on-semi-self-and-unsupervised-learning-for-image-543k9cruy6.pdf

A Survey on Semi-, Self- and Unsupervised Learning for Image Classification

Cloud computing is extensively used in a variety of applications and domains, however task and resource scheduling remains an area that requires improvement. Put simply, in a heterogeneous computing system, task scheduling algorithms, which allow the transfer of incoming tasks to machines, are needed to satisfy high performance data mapping requirements. The appropriate mapping between resources and tasks reduces makespan and maximises resource utilisation. In this contribution, we present a novel scheduling algorithm using Directed Acyclic Graph (DAG) based on the Prediction of Tasks Computation Time algorithm (PTCT) to estimate the preeminent scheduling algorithm for prominent cloud data. In addition, the proposed algorithm provides a significant improvement with respect to the makespan and reduces the computation and complexity via employing Principle Components Analysis (PCA) and reducing the Expected Time to Compute (ETC) matrix. Simulation results confirm the superior performance of the algorithm for heterogeneous systems in terms of efficiency, speedup and schedule length ratio, when compared to the state-of-the-art Min-Min, Max-Min, QoS-Guide and MiM-MaM scheduling algorithms.

/pdf/a-task-scheduling-algorithm-with-improved-makespan-based-on-20c2veznco.pdf

A Task Scheduling Algorithm With Improved Makespan Based on Prediction of Tasks Computation Time algorithm for Cloud Computing

Exact solution techniques for two-dimensional cutting and packing

Image-to-image translation (I2I) aims to transfer images from a source domain to a target domain while preserving the content representations. I2I has drawn increasing attention and made tremendous progress in recent years because of its wide range of applications in many computer vision and image processing problems, such as image synthesis, segmentation, style transfer, restoration, and pose estimation. In this paper, we provide an overview of the I2I works developed in recent years. We will analyze the key techniques of the existing I2I works and clarify the main progress the community has made. Additionally, we will elaborate on the effect of I2I on the research and industry community and point out remaining challenges in related fields. 

Image-to-Image Translation: Methods and Applications

Deep neural networks are typically trained under a supervised learning framework where a model learns a single task using labeled data. Instead of relying solely on labeled data, practitioners can harness unlabeled or related data to improve model performance, which is often more accessible and ubiquitous. Self-supervised pre-training for transfer learning is becoming an increasingly popular technique to improve state-of-the-art results using unlabeled data. It involves first pre-training a model on a large amount of unlabeled data, then adapting the model to target tasks of interest. In this review, we survey self-supervised learning methods and their applications within the sequential transfer learning framework. We provide an overview of the taxonomy for self-supervised learning and transfer learning, and highlight some prominent methods for designing pre-training tasks across different domains. Finally, we discuss recent trends and suggest areas for future investigation.

A Survey on Self-supervised Pre-training for Sequential Transfer Learning in Neural Networks.

While deep learning strategies achieve outstanding results in computer vision tasks, one issue remains: The current strategies rely heavily on a huge amount of labeled data. In many real-world problems, it is not feasible to create such an amount of labeled training data. Therefore, it is common to incorporate unlabeled data into the training process to reach equal results with fewer labels. Due to a lot of concurrent research, it is difficult to keep track of recent developments. In this survey, we provide an overview of often used ideas and methods in image classification with fewer labels. We compare 25 methods in detail. In our analysis, we identify three major trends. 1. State-of-theart methods are scaleable to real-world applications based on their accuracy. 2. The degree of supervision which is needed to achieve comparable results to the usage of all labels is decreasing. 3. All methods share common ideas while only a few methods combine these ideas to achieve better performance. Based on all of these three trends we discover future research opportunities.

A survey on Semi-, Self- and Unsupervised Techniques in Image Classification.

We study the Parallel Task Scheduling problem $Pm|size_j|C_{\max}$ with a constant number of machines. This problem is known to be strongly NP-complete for each $m \geq 5$, while it is solvable in pseudo-polynomial time for each $m \leq 3$. We give a positive answer to the long-standing open question whether this problem is strongly $NP$-complete for $m=4$. As a second result, we improve the lower bound of $\frac{12}{11}$ for approximating pseudo-polynomial Strip Packing to $\frac{5}{4}$. Since the best known approximation algorithm for this problem has a ratio of $\frac{4}{3} + \varepsilon$, this result narrows the gap between approximation ratio and inapproximability result by a significant step. Both results are proven by a reduction from the strongly $NP$-complete problem 3-Partition.

Complexity and Inapproximability Results for Parallel Task Scheduling and Strip Packing

Collagen fiber orientations in bones, visible with Second Harmonic Generation (SHG) microscopy, represent the inner structure and its alteration due to influences like cancer. While analyses of these orientations are valuable for medical research, it is not feasible to analyze the needed large amounts of local orientations manually. Since we have uncertain borders for these local orientations only rough regions can be segmented instead of a pixel-wise segmentation. We analyze the effect of these uncertain borders on human performance by a user study. Furthermore, we compare a variety of 2D and 3D methods such as classical approaches like Fourier analysis with state-of-the-art deep neural networks for the classification of local fiber orientations. We present a general way to use pretrained 2D weights in 3D neural networks, such as Inception-ResNet-3D a 3D extension of Inception-ResNet-v2. In a 10 fold cross-validation our two stage segmentation based on Inception-ResNet-3D and transferred 2D ImageNet weights achieves a human comparable accuracy.

2D and 3D Segmentation of Uncertain Local Collagen Fiber Orientations in SHG Microscopy

We study Parallel Task Scheduling $Pm|size_j|C_{\max }$ with a constant number of machines. This problem is known to be strongly NP-complete for each $m \ge 5$, while it is solvable in pseudo-polynomial time for each $m \le 3$. We give a positive answer to the long-standing open question whether this problem is strongly NP-complete for $m=4$. As a second result, we improve the lower bound of $\frac{12}{11}$ for approximating pseudo-polynomial Strip Packing to $\frac{5}{4}$. Since the best known approximation algorithm for this problem has a ratio of $\frac{4}{3} + \varepsilon $, this result narrows the gap between approximation ratio and inapproximability result by a significant step. Both results are proven by a reduction from the strongly NP-complete problem 3-Partition.

Lars Schmarje

Papers

A Survey on Semi-, Self- and Unsupervised Learning for Image Classification

A survey on Semi-, Self- and Unsupervised Techniques in Image Classification.

Complexity and Inapproximability Results for Parallel Task Scheduling and Strip Packing

2D and 3D Segmentation of Uncertain Local Collagen Fiber Orientations in SHG Microscopy

Complexity and Inapproximability Results for Parallel Task Scheduling and Strip Packing