Is it Enough to Optimize CNN Architectures on ImageNet?

• URL: http://arxiv.org/abs/2103.09108v1
• Date: Tue, 16 Mar 2021 14:42:01 GMT
• Title: Is it Enough to Optimize CNN Architectures on ImageNet?
• Authors: Lukas Tuggener, J\"urgen Schmidhuber, Thilo Stadelmann
• Abstract summary: We train 500 CNN architectures on ImageNet and 8 other image classification datasets. The relationship between architecture and performance varies wildly, depending on the datasets. We identify two dataset-specific performance indicators: the cumulative width across layers as well as the total depth of the network.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: An implicit but pervasive hypothesis of modern computer vision research is that convolutional neural network (CNN) architectures that perform better on ImageNet will also perform better on other vision datasets. We challenge this hypothesis through an extensive empirical study for which we train 500 sampled CNN architectures on ImageNet as well as 8 other image classification datasets from a wide array of application domains. The relationship between architecture and performance varies wildly, depending on the datasets. For some of them, the performance correlation with ImageNet is even negative. Clearly, it is not enough to optimize architectures solely for ImageNet when aiming for progress that is relevant for all applications. Therefore, we identify two dataset-specific performance indicators: the cumulative width across layers as well as the total depth of the network. Lastly, we show that the range of dataset variability covered by ImageNet can be significantly extended by adding ImageNet subsets restricted to few classes.

Related papers

• ImageNot: A contrast with ImageNet preserves model rankings [16.169858780154893]
  We introduce ImageNot, a dataset designed to match the scale of ImageNet while differing drastically in other aspects. Key model architectures developed for ImageNet over the years rank identically when trained and evaluated on ImageNot to how they rank on ImageNet.
  arXiv  Detail & Related papers  (2024-04-02T17:13:04Z)
• Does progress on ImageNet transfer to real-world datasets? [28.918770106968843]
  We evaluate ImageNet pre-trained models with varying accuracy on six practical image classification datasets. On multiple datasets, models with higher ImageNet accuracy do not consistently yield performance improvements. We hope that future benchmarks will include more diverse datasets to encourage a more comprehensive approach to improving learning algorithms.
  arXiv  Detail & Related papers  (2023-01-11T18:55:53Z)
• BigDatasetGAN: Synthesizing ImageNet with Pixel-wise Annotations [89.42397034542189]
  We synthesize a large labeled dataset via a generative adversarial network (GAN) We take image samples from the class-conditional generative model BigGAN trained on ImageNet, and manually annotate 5 images per class, for all 1k classes. We create a new ImageNet benchmark by labeling an additional set of 8k real images and evaluate segmentation performance in a variety of settings.
  arXiv  Detail & Related papers  (2022-01-12T20:28:34Z)
• KiU-Net: Overcomplete Convolutional Architectures for Biomedical Image and Volumetric Segmentation [71.79090083883403]
  "Traditional" encoder-decoder based approaches perform poorly in detecting smaller structures and are unable to segment boundary regions precisely. We propose KiU-Net which has two branches: (1) an overcomplete convolutional network Kite-Net which learns to capture fine details and accurate edges of the input, and (2) U-Net which learns high level features. The proposed method achieves a better performance as compared to all the recent methods with an additional benefit of fewer parameters and faster convergence.
  arXiv  Detail & Related papers  (2020-10-04T19:23:33Z)
• Dynamic Graph: Learning Instance-aware Connectivity for Neural Networks [78.65792427542672]
  Dynamic Graph Network (DG-Net) is a complete directed acyclic graph, where the nodes represent convolutional blocks and the edges represent connection paths. Instead of using the same path of the network, DG-Net aggregates features dynamically in each node, which allows the network to have more representation ability.
  arXiv  Detail & Related papers  (2020-10-02T16:50:26Z)
• U-Net Based Architecture for an Improved Multiresolution Segmentation in Medical Images [0.0]
  We have proposed a fully convolutional neural network for image segmentation in a multi-resolution framework. In the proposed architecture (mrU-Net), the input image and its down-sampled versions were used as the network inputs. We trained and tested the network on four different medical datasets.
  arXiv  Detail & Related papers  (2020-07-16T10:19:01Z)
• From ImageNet to Image Classification: Contextualizing Progress on Benchmarks [99.19183528305598]
  We study how specific design choices in the ImageNet creation process impact the fidelity of the resulting dataset. Our analysis pinpoints how a noisy data collection pipeline can lead to a systematic misalignment between the resulting benchmark and the real-world task it serves as a proxy for.
  arXiv  Detail & Related papers  (2020-05-22T17:39:16Z)
• Medical Image Segmentation Using a U-Net type of Architecture [0.0]
  We argue that the architecture of U-Net, when combined with a supervised training strategy at the bottleneck layer, can produce comparable results with the original U-Net architecture. We introduce a fully supervised FC layers based pixel-wise loss at the bottleneck of the encoder branch of U-Net. The two layer based FC sub-net will train the bottleneck representation to contain more semantic information, which will be used by the decoder layers to predict the final segmentation map.
  arXiv  Detail & Related papers  (2020-05-11T16:10:18Z)
• Improved Residual Networks for Image and Video Recognition [98.10703825716142]
  Residual networks (ResNets) represent a powerful type of convolutional neural network (CNN) architecture. We show consistent improvements in accuracy and learning convergence over the baseline. Our proposed approach allows us to train extremely deep networks, while the baseline shows severe optimization issues.
  arXiv  Detail & Related papers  (2020-04-10T11:09:50Z)
• CRNet: Cross-Reference Networks for Few-Shot Segmentation [59.85183776573642]
  Few-shot segmentation aims to learn a segmentation model that can be generalized to novel classes with only a few training images. With a cross-reference mechanism, our network can better find the co-occurrent objects in the two images. Experiments on the PASCAL VOC 2012 dataset show that our network achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2020-03-24T04:55:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.