Anomaly Detection on Fishyscapes L&F

Metric: FPR95 (higher is better)

LeaderboardDataset

Loading chart...

Results

Hide augmentations

Sort:

#	Model↕	FPR95▼	Augmentations	Paper	Date↕	Code
1	Dirichlet DeepLab	47.43	Yes	The Fishyscapes Benchmark: Measuring Blind Spots...	2019-04-05	Code
2	Softmax Entropy	44.8	No	The Fishyscapes Benchmark: Measuring Blind Spots...	2019-04-05	Code
3	Bayesian DeepLab	38.5	No	The Fishyscapes Benchmark: Measuring Blind Spots...	2019-04-05	Code
4	Learned Embedding Density	24.4	Yes	The Fishyscapes Benchmark: Measuring Blind Spots...	2019-04-05	Code
5	Void Classifier	22.11	Yes	The Fishyscapes Benchmark: Measuring Blind Spots...	2019-04-05	Code
6	OutlierHead combined instances	19.02	No	Simultaneous Semantic Segmentation and Outlier D...	2019-08-03	Code
7	SynBoost	15.79	Yes	Pixel-wise Anomaly Detection in Complex Driving ...	2021-03-09	Code
8	SML	14.53	No	Standardized Max Logits: A Simple yet Effective ...	2021-07-23	Code
9	CosMe	13.32	No	Consensus Synergizes with Memory: A Simple Appro...	2021-11-24	-
10	cDNP	8.9	No	Far Away in the Deep Space: Dense Nearest-Neighb...	2022-11-12	Code
11	PEBAL	7.58	Yes	Pixel-wise Energy-biased Abstention Learning for...	2021-11-24	Code
12	cDNP+OE	7.5	Yes	Far Away in the Deep Space: Dense Nearest-Neighb...	2022-11-12	Code
13	DenseHybrid	6.2	Yes	DenseHybrid: Hybrid Anomaly Detection for Dense ...	2022-07-06	Code
14	FlowEneDet	5.2	No	Concurrent Misclassification and Out-of-Distribu...	2023-05-16	Code
15	Mask2Anomaly	4.36	Yes	Unmasking Anomalies in Road-Scene Segmentation	2023-07-25	Code
16	RPL+CoroCL	2.27	Yes	Residual Pattern Learning for Pixel-wise Out-of-...	2022-11-26	Code
17	NFlowJS-GF (with extra inlier set: Vistas and Wilddash2)	2	Yes	Dense Out-of-Distribution Detection by Robust Le...	2021-12-23	-

#1Dirichlet DeepLabSOTA
47.43
FPR95· Augmentations· 2019-04-05
The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation Code
#2Softmax Entropy
44.8
FPR95· 2019-04-05
The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation Code
#3Bayesian DeepLab
38.5
FPR95· 2019-04-05
The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation Code
#4Learned Embedding Density
24.4
FPR95· Augmentations· 2019-04-05
The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation Code
#5Void Classifier
22.11
FPR95· Augmentations· 2019-04-05
The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation Code
#6OutlierHead combined instances
19.02
FPR95· 2019-08-03
Simultaneous Semantic Segmentation and Outlier Detection in Presence of Domain Shift Code
#7SynBoost
15.79
FPR95· Augmentations· 2021-03-09
Pixel-wise Anomaly Detection in Complex Driving Scenes Code
#8SML
14.53
FPR95· 2021-07-23
Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentation Code
#9CosMe
13.32
FPR95· 2021-11-24
Consensus Synergizes with Memory: A Simple Approach for Anomaly Segmentation in Urban Scenes
#10cDNP
8.9
FPR95· 2022-11-12
Far Away in the Deep Space: Dense Nearest-Neighbor-Based Out-of-Distribution Detection Code
#11PEBAL
7.58
FPR95· Augmentations· 2021-11-24
Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes Code
#12cDNP+OE
7.5
FPR95· Augmentations· 2022-11-12
Far Away in the Deep Space: Dense Nearest-Neighbor-Based Out-of-Distribution Detection Code
#13DenseHybrid
6.2
FPR95· Augmentations· 2022-07-06
DenseHybrid: Hybrid Anomaly Detection for Dense Open-set Recognition Code
#14FlowEneDet
5.2
FPR95· 2023-05-16
Concurrent Misclassification and Out-of-Distribution Detection for Semantic Segmentation via Energy-Based Normalizing Flow Code
#15Mask2Anomaly
4.36
FPR95· Augmentations· 2023-07-25
Unmasking Anomalies in Road-Scene Segmentation Code
#16RPL+CoroCL
2.27
FPR95· Augmentations· 2022-11-26
Residual Pattern Learning for Pixel-wise Out-of-Distribution Detection in Semantic Segmentation Code
#17NFlowJS-GF (with extra inlier set: Vistas and Wilddash2)
2
FPR95· Augmentations· 2021-12-23
Dense Out-of-Distribution Detection by Robust Learning on Synthetic Negative Data