AI summaryⓘ
The authors point out that object detection systems work well in normal conditions but can fail without warning when images are blurry or noisy. They developed a method that helps detectors recognize when an image's quality has dropped by organizing image features based on types and levels of degradation rather than the object content. This approach uses a special learning technique to group similar degraded images together, creating a measure of how far input images diverge from clean ones. Their experiments show this method reliably detects image quality issues across different detectors and conditions without needing extra labels. This helps object detectors internally know when their inputs are outside typical quality ranges, improving safety and reliability.
object detectionimage degradationcontrastive learningfeature spacedegradation manifoldzero-shot transferdistribution shiftembeddingpristine prototypeself-awareness
Authors
Stefan Becker, Simon Weiss, Wolfgang Hübner, Michael Arens
Abstract
Object detectors achieve strong performance under nominal imaging conditions but can fail silently when exposed to blur, noise, compression, adverse weather, or resolution changes. In safety-critical settings, it is therefore insufficient to produce predictions without assessing whether the input remains within the detector's nominal operating regime. We refer to this capability as self-aware object detection. We introduce a degradation-aware self-awareness framework based on degradation manifolds, which explicitly structure a detector's feature space according to image degradation rather than semantic content. Our method augments a standard detection backbone with a lightweight embedding head trained via multi-layer contrastive learning. Images sharing the same degradation composition are pulled together, while differing degradation configurations are pushed apart, yielding a geometrically organized representation that captures degradation type and severity without requiring degradation labels or explicit density modeling. To anchor the learned geometry, we estimate a pristine prototype from clean training embeddings, defining a nominal operating point in representation space. Self-awareness emerges as geometric deviation from this reference, providing an intrinsic, image-level signal of degradation-induced shift that is independent of detection confidence. Extensive experiments on synthetic corruption benchmarks, cross-dataset zero-shot transfer, and natural weather-induced distribution shifts demonstrate strong pristine-degraded separability, consistent behavior across multiple detector architectures, and robust generalization under semantic shift. These results suggest that degradation-aware representation geometry provides a practical and detector-agnostic foundation.