Few-shot object detection, the problem of modelling novel object detection categories with few training instances, is an emerging topic in the area of few-shot learning and object detection. Contemporary techniques can be divided into two groups: fine-tuning based and meta-learning based approaches. While meta-learning approaches aim to learn dedicated meta-models for mapping samples to novel class models, fine-tuning approaches tackle few-shot detection in a simpler manner, by adapting the detection model to novel classes through gradient based optimization. Despite their simplicity, fine-tuning based approaches typically yield competitive detection results. Based on this observation, we focus on the role of loss functions and augmentations as the force driving the fine-tuning process, and propose to tune their dynamics through meta-learning principles. The proposed training scheme, therefore, allows learning inductive biases that can boost few-shot detection, while keeping the advantages of fine-tuning based approaches. In addition, the proposed approach yields interpretable loss functions, as opposed to highly parametric and complex few-shot meta-models. The experimental results highlight the merits of the proposed scheme, with significant improvements over the strong fine-tuning based few-shot detection baselines on benchmark Pascal VOC and MS-COCO datasets, in terms of both standard and generalized few-shot performance metrics.

In this work, we propose a zero-shot learning method to effectively model knowledge transfer between classes via jointly learning visually consistent word vectors and label embedding model in an end-to-end manner. The main idea is to project the vector space word vectors of attributes and classes into the visual space such that word representations of semantically related classes become more closer, and use the projected vectors in the proposed embedding model to identify unseen classes. We evaluate the proposed approach on two benchmark datasets and the experimental results show that our method yields significant improvements in recognition accuracy.

In this paper, the hyperspectral detection of targets in visible-near infrared (VNIR) images is studied. The change of radiance domain signatures in images taken in different locations, time and altitudes are analyzed. A new radiance domain detection scheme for VNIR images under 1000 m altitude is proposed. The analysis shows that the radiance domain signatures of each target, that are collected from an image taken at 10 m altitude, can be effectively used for pure pixel target detection in other VNIR images taken at altitudes between 10 - 1000 m . The proposed approach is tested using several target types and on images taken at different altitudes and environmental conditions. Our results show that target detection in radiance domain provides a cheaper, easier and effective alternative to reflectance domain, in VNIR images.

In this study, the effect of utilizing different type of signatures on plant detection success is evaluated on hyperspectral aerial images. Plant regions are tried to detect using spectral signatures of leaf, stem and tassel belonging to the plant separately and the plant representative signature (PRS) is created by averaging of signatures of selected region on the aerial images. The signatures used for detection are generated from hyperspectral images taken from 10m distance to target plant. The Spectral Angle Mapper(SAM) and Generalized Likelihood Ratio Test(G Lrt)algorithms are used for target detection. Performance evaluation is made by Receiver Operating Characteristic (ROC) curves. When the results are evaluated, it is observed that the detection performance with the use of PRS is higher.

We propose a novel approach for unsupervised zero-shot learning (ZSL) of classes based on their names. Most existing unsupervised ZSL methods aim to learn a model for directly comparing image features and class names. However, this proves to be a difficult task due to dominance of non-visual semantics in underlying vector-space embeddings of class names. To address this issue, we discriminatively learn a word representation such that the similarities between class and combination of attribute names fall in line with the visual similarity. Contrary to the traditional zero-shot learning approaches that are built upon attribute presence, our approach bypasses the laborious attribute-class relation annotations for unseen classes. In addition, our proposed approach renders text-only training possible, hence, the training can be augmented without the need to collect additional image data. The experimental results show that our method yields state-of-the-art results for unsupervised ZSL in three benchmark datasets.