Pith. sign in

REVIEW 9 cited by

FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1612.02649 v1 pith:7FIC3POL submitted 2016-12-08 cs.CV

Judy Hoffman , Dequan Wang , Fisher Yu , Trevor Darrell This is my paper
classification cs.CV
keywords domainadaptationadversarialdifferentacrossapproachcategorycity
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Fully convolutional models for dense prediction have proven successful for a wide range of visual tasks. Such models perform well in a supervised setting, but performance can be surprisingly poor under domain shifts that appear mild to a human observer. For example, training on one city and testing on another in a different geographic region and/or weather condition may result in significantly degraded performance due to pixel-level distribution shift. In this paper, we introduce the first domain adaptive semantic segmentation method, proposing an unsupervised adversarial approach to pixel prediction problems. Our method consists of both global and category specific adaptation techniques. Global domain alignment is performed using a novel semantic segmentation network with fully convolutional domain adversarial learning. This initially adapted space then enables category specific adaptation through a generalization of constrained weak learning, with explicit transfer of the spatial layout from the source to the target domains. Our approach outperforms baselines across different settings on multiple large-scale datasets, including adapting across various real city environments, different synthetic sub-domains, from simulated to real environments, and on a novel large-scale dash-cam dataset.

Discussion (0). Sign in to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Blending-target Domain Adaptation by Adversarial Meta-Adaptation Networks

    cs.LG 2019-07 unverdicted novelty 7.0 of 10

    AMEAN applies adversarial meta-learning to discover implicit meta-sub-target clusters in blended target data, reducing intra-target category misalignment and outperforming standard DA methods on three BTDA benchmarks.

  2. PairedGTA: Generating Driving Datasets for Controlled Photometric Shift Analysis

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    PairedGTA uses a game engine to produce pixel-aligned driving images under controlled photometric variations for isolated evaluation of environmental effects on perception models.

  3. Make me an Expert: Distilling from Generalist Black-Box Models into Specialized Models for Semantic Segmentation

    cs.CV 2025-08 conditional novelty 6.0 of 10

    ATGC selects the best input scale for a black-box open-vocabulary segmentation API, using DINOv2 attention entropy, improving one-hot-label distillation on Cityscapes and ACDC.

  4. Boundary and Entropy-driven Adversarial Learning for Fundus Image Segmentation

    cs.CV 2019-06 unverdicted novelty 6.0 of 10

    BEAL improves cross-domain optic disc and cup segmentation by adversarially aligning boundary predictions and entropy maps from source to target domains.

  5. Interact3D: Compositional 3D Generation of Interactive Objects

    cs.CV 2026-03 unverdicted novelty 5.0 of 10

    Interact3D composes multi-object 3D scenes from one image via asset generation, registration plus SDF collision penalties, and VLM-driven closed-loop image edits.

  6. $\varphi$-Adapt: A Physics-Informed Adaptation Learning Approach to 2D Quantum Material Discovery

    cs.CV 2025-07 reject novelty 5.0 of 10

    A physics-informed domain adaptation approach, trained on 600,000 synthesized flake images, is claimed to set state-of-the-art results for detection, layer classification, and thickness estimation on real 2D material ...

  7. Real-Time Per-Garment Virtual Try-On with Temporal Consistency for Loose-Fitting Garments

    cs.GR 2025-06 conditional novelty 5.0 of 10

    A per-garment virtual try-on method for loose-fitting garments uses a garment-invariant pose representation and a recurrent ConvLSTM synthesis network to achieve temporally smoother try-on video at about 10 fps.

  8. TITAN: Query-Token based Domain Adaptive Adversarial Learning

    cs.CV 2025-06 reject novelty 4.0 of 10

    TITAN claims large source-free domain adaptation gains using variance-based target splitting and query-token adversarial alignment, but internal contradictions and test-set leakage invalidate the reported results.

  9. How much real data do we actually need: Analyzing object detection performance using synthetic and real data

    cs.CV 2019-07 unverdicted novelty 3.0 of 10

    Synthetic data can partially substitute for real data in object detection training, with performance tied to domain similarity and the volume of real data included.

Pith tools