We address the problem of open-vocabulary object retrieval and localization, which is to retrieve and localize objects from a very large-scale image database immediately by a textual query (e.g., a word or phrase). We first propose Query-Adaptive R-CNN, a simple yet strong framework for open-vocabulary object detection. Query-Adaptive RCNN is a simple extension of Faster R-CNN from closedvocabu...

We have analyzed the measurements of end-to-end VoIP packets by tracing UDP packets between 12 testboxes in the Netherlands. We show that voice probes experience low delay and loss in the network. Moreover, we find that the reordering of packets has virtually no impact on the voice quality in our experiments. To determine the quality of VoIP calls over time, we have monitored the end-to-end pac...

Understanding the service quality of Internet paths is of great importance for both the service provider and the user. Motivated by this and particularly the aim of extracting such information from a given dataset collected through end-to-end measurement, this paper presents a novel approach, called SCI (Service-quality Characterization of Internet-path). Specifically, SCI utilizes delay and lo...

We introduce an efficient protocol for end-to-end handoff management in heterogeneous wireless IP-based networks. The protocol is based on the stream control transmission protocol (SCTP), and employs a soft-handoff mechanism that uses end-to-end semantics for signaling handoffs and for transmitting control messages. The design goal of this protocol is twofold—first, to reduce the home registrat...

Current best local descriptors are learned on a large dataset of matching and non-matching keypoint pairs. However, data of this kind is not always available since detailed keypoint correspondences can be hard to establish. On the other hand, we can often obtain labels for pairs of keypoint bags. For example, keypoint bags extracted from two images of the same object under different views form ...

The mobile wireless environment can be challenging for multimedia applications, with poor performance and handover issues causing service degradation. In this paper, a modified version of the new IETF transport protocol, SCTP, is shown to improve the performance of a voice-like application in an environment where two wireless LAN’s overlap. It does this by monitoring both networks and selecting...

In this paper, we study the task of detecting semantic parts of an object, e.g., a wheel of a car, under partial occlusion. We propose that all models should be trained without seeing occlusions while being able to transfer the learned knowledge to deal with occlusions. This setting alleviates the difficulty in collecting an exponentially large dataset to cover occlusion patterns and is more es...

Weakly supervised object detection (WSOD), which is the problem of learning detectors using only image-level labels, has been attracting more and more interest. However, this problem is quite challenging due to the lack of location supervision. To address this issue, this paper integrates saliency into a deep architecture, in which the location information is explored both explicitly and implic...

Self-stabilizing protocols must begin operating correctly even when started from an arbitrary state. The end-to-end problem is to ensure reliable message delivery across an unreliable network under the weakest possible guarantee from the network { that the sender and receiver are never separated by a cut of permanently failed links. In this paper we present the rst self-stabilizing end-to-end p...