Abstract:
The overall architecture and details of a scalable video fingerprinting and identification system that is robust with respect to many classes of video distortions is described. In this system, a fingerprint for a piece of multimedia content is composed of a number of compact signatures, along with traversal hash signatures and associated metadata. Numerical descriptors are generated for features found in a multimedia clip, signatures are generated from these descriptors, and a reference signature database is constructed from these signatures. Query signatures are also generated for a query multimedia clip. These query signatures are searched against the reference database using a fast similarity search procedure, to produce a candidate list of matching signatures. This candidate list is further analyzed to find the most likely reference matches. Signature correlation is performed between the likely reference matches and the query clip to improve detection accuracy.
Abstract:
A method is presented for large media data base query and media entry identification based on multi-level similarity search and reference-query entry correlation. Media content fingerprinting detects unique features and generates discriminative descriptors and signatures used to form preliminary reference data base. The preliminary reference data base is processed and a subset-set of it is selected to form a final reference data base. To identify a media query a fast similarity search is performed first on the reference database resulting in a preliminary set of likely matching videos. For each preliminary likely matching video a further multi-level correlation is performed which includes iterative refinement, sub-sequence merging, and final result classification.
Abstract:
The overall architecture and details of a scalable video fingerprinting and identification system that is robust with respect to many classes of video distortions is described. In this system, a fingerprint for a piece of multimedia content is composed of a number of compact signatures, along with traversal hash signatures and associated metadata. Numerical descriptors are generated for features found in a multimedia clip, signatures are generated from these descriptors, and a reference signature database is constructed from these signatures. Query signatures are also generated for a query multimedia clip. These query signatures are searched against the reference database using a fast similarity search procedure, to produce a candidate list of matching signatures. This candidate list is further analyzed to find the most likely reference matches. Signature correlation is performed between the likely reference matches and the query clip to improve detection accuracy.
Abstract:
A method is presented for large media data base query and media entry identification based on multi-level similarity search and reference-query entry correlation. Media content fingerprinting detects unique features and generates discriminative descriptors and signatures used to form preliminary reference data base. The preliminary reference data base is processed and a subset-set of it is selected to form a final reference data base. To identify a media query a fast similarity search is performed first on the reference database resulting in a preliminary set of likely matching videos. For each preliminary likely matching video a further multi-level correlation is performed which includes iterative refinement, sub-sequence merging, and final result classification.
Abstract:
Multiple network switches are configured having memory interfaces that transfer segmented frame data to each other via a data path having multiple rings connecting the network switches. The memory interfaces are also configured for transferring the segmented frame data to respective local buffer memories for temporary storage. The data path transfers the data units between the switches according to a prescribed sequence protocol, optimizing memory bandwidth by requiring only one read and one write operation to and from a local buffer memory for each segmented frame data being received and transmitted through the switches.
Abstract:
A network switch configured for switching data packets across multiple ports uses decision making logic to generate frame forwarding information. The decision making logic employs a pipelined architecture that enables multiple data frames to be processed simultaneously to increase data throughput. The decision making logic also pipelines access to an address lookup table that stores the data forwarding information. An arbitration circuit provides the decision making device with automatic access to the address table in alternate time slots and also enables other circuits to access the address table in predetermined time slots.
Abstract:
A novel method of enabling a port of a network switch to support connections with multiple VLANs. The method comprises storing VLAN data indicating a plurality of VLAN identifiers corresponding to the multiple VLANs supported by the port. A VLAN identifier of a data packet received via the port is compared with the plurality of VLAN identifiers determined using the stored VLAN data. The data packet is forwarded for further processing if the VLAN identifier matches one of the plurality of VLAN identifiers. However, the data packet is discarded if the VLAN identifier does not match one of the plurality of VLAN identifiers. Moreover, VLAN information corresponding to a VLAN identifier of a data packet to be transmitted from the port is compared with the stored VLAN data to determine whether the VLAN identifier matches one of the plurality of VLAN identifiers supported by the port. The data packet is prevented from being transmitted from the port if the VLAN identifier does not match one of the plurality of VLAN identifiers.
Abstract:
Multiple network switch modules have memory interfaces configured for transferring packet data to respective buffer memories. The memory interfaces are also configured for transfer among each other data units of data frames received from different network switch modules. A shared switching logic monitors (“snoops”) the data units as they are transferred between network switch modules, providing a centralized switching decision logic for multiple network switch modules. The memory interfaces transfer the data units according to a prescribed sequence, optimizing memory bandwidth by concurrently executing a prescribed number of successive memory writes or memory reads. A preferred embodiment includes a distributed memory interface in between the network switch modules and a shared memory system.
Abstract:
A network switch configured for switching data packets across multiple ports uses an address table to generate frame forwarding information. The address table includes aging information used by an internal decision making engine so that aged entries are invalidated after a programmable period of time. The network switch includes a test mode so that the aging function is able to be quickly tested by an external test device.
Abstract:
A network switch having switch ports for full-duplex communication of data packets with respective network nodes according to Ethernet (IEEE 802.3) protocol dynamically allocates external memory bandwidth slots between high data rate ports. An external memory interface determines if a high data rate port makes a request for a bandwidth slot and grants the request if made. The slot is taken from a selected group which is a subset of the total number of slots. If a request for the slot is not made, the external memory interface assigns the slot to another high data rate port. Lower data rate ports in the network switch are assigned fixed slots from those slots not from within the selected group of slots. The dynamic allocation of bandwidth slots between the high data rate port enables the efficient use of limited memory bandwidth resources.