AI Paper Atlas

Research map / Detection and segmentation / CNN features

HexNet: An Orientation-Aware Deep Learning Framework for Omni-Directional Input

Chao Zhang, Stephan Liwicki, Sen He, William A. P. Smith, Roberto Cipolla

TPAMI 2023 · Detection and segmentation › CNN features · approach: Convolution-transformer backbones · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add arc-based north-aligned interpolation (ablation)Hex-UNet · Synthia-O (r=7)mIoU: 45.1 -> 49.5Improved
initialise from perspective U-Net filters with 10 epochs refinement (Hex-UNet-T) instead of training from scratchHex-UNet · Synthia-O (r=6)mIoU: 43.6 -> 36.7Hurt
filter transfer at higher resolutionHex-UNet-T · Synthia-O (r=8)mIoU: 45.3 vs 47.1Mixed
use spherical instead of planar representationU-Net · Synthia-S (lv-8 vs planar 1/3)mIoU: 56.5 -> 57.6Improved
use spherical instead of planar representationU-Net · Cityscapes (1/6 res)mIoU: 51.5 -> 54.3Improved
use spherical representation with ImageNet pretrainingDANet · Cityscapes (1/6 res)mIoU: 52.5 -> 56.6Improved
add ImageNet pretraining to planar DANetDANet · Cityscapes (1/6 res)mIoU: 51.2 -> 57.9Improved
replace 3x3 kernel with hexagonal kernel (no sphere) (ablation)U-Net · Cityscapes (1/3 res)mIoU: 55.5 -> 55.2No clear effect
replace 3x3 kernel with hexagonal kernel (no sphere), pretrained (ablation)DANet · Cityscapes (1/6 res)mIoU: 57.9 -> 57.1Hurt
raise mesh resolution from r=6 to r=7 (ablation)Hex-UNet · Synthia-OmIoU: 43.6 -> 49.5Improved
raise mesh resolution from r=7 to r=8 (ablation)Hex-UNet · Synthia-OmIoU: 49.5 -> 47.1Hurt
use masked partial-sphere computation (ablation)Graph-HexNet · Synthia-O (r=7)memory (MB): 3,596 -> 595Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy