About TNT: A Brief Technical Overview

At the core of TNT is an extremely powerful computing system. TNT runs the image through an advanced image processing algorithm, which extracts the significant features associated with the species and identifies the individual markings for that animal. Those are then compared against the markings of other animals of the same species contained in a database that encompasses the whole world. We can process an image and determine in milliseconds if that image is of a previously ‘tagged’ animal.

The only requirements for using TNT are a digital camera and an internet connection. The power of TNT is however, not in the technology of image processing and pattern recognition – it is in the power of the collection of information by an ‘army’ of interested individuals and storing that information in a database that is created by the work of people all over the world and available to everyone in the world.


TNT addresses the three main problem areas associated with traditional tagging methods: 1) Scalability, 2) labor-intensiveness, and 3) mutilation of subjects. With TNT, we like to frame those considerations in terms of Better, Cheaper, and Faster.

TNT is not a hypothetical technology that only works in a lab setting. TNT was conceived of all the way back in 1985. However, the technology of high speed computers, digital cameras, and especially The Web, were some 20 years off. The time has now come.


We have developed digital image processing and statistical pattern recognition methods that allow repeated, unambiguous identification of individual animals. This means that, except in the rarest of cases, there is no longer a need for invasive or mutilation tagging methods. This eliminates any scientific concern regarding the effects of the tag on the subject animal’s behavior and any legal or regulatory issues regarding humane or ethical treatment of research animals.

A natural tag – one that is part of the animal – cannot be lost. With some species, traditional tagging methods are problematic. Amphibians, for example, (a) when toe-clipped often regenerate the toe; or, with PIT (Passive Integrated Transponder) tags (b) are easily infected by, or (c) may expel them.

Not having to deal with tag loss can substantially simplify the math associated with population estimates, while at the same time increasing the statistical power of the estimate.

The only limit on the tag numbers available is the number of animals in the population. There is no requirement to register your code, as with thermal marking in fish. It's not necessary to coordinate codes as with coded wire tags. There is no need for a series or "reservation" for a unique series of tag numbers.

Morphometric data can be extracted automatically from image data. Length, gender and condition algorithms have been established for the test species we have examined. Geographic (GPS) data is easily added to the individual’s record. Whatever ancillary data the researcher wants to add can be done automatically.

Of course, TNT cannot replace radio tags, used when one wishes to track the local movement of an individual. Nor can it replace satellite tags where unmanned, quasi-real-time tracking over large geographic distances is required.


Part of TNT's business development required examination of the traditional methods used in population estimates and their associated costs. In all cases, TNT is less expensive. In the case of coded wire tags, TNT's cost per-returned-tag is approximately 80% less. In comparison to PIT tags and t-bar tags, TNT cost is half or less.

Among the factors that drive reduced expense: TNT eliminates the costs of tags and tag application equipment. It also eliminates administrative overhead associated with keeping and verifying unique tag numbers.

In most cases, field biologists currently involved with population assessment will change their methods only slightly – take pictures instead of cut, poke, tattoo, brand, etc. In the cases we have examined so far, TNT implementation results in a reduction of effort.


Regardless of species, standard turn-around time for a complete report is a week. If urgent, this time can be reduced to as little as three days. There are qualifiers of course. If, for example, processing half-a-million juvenile salmon images from a hatchery, turnaround time will be longer. However, when processing smolt images for example, the issue is not ID, but input, and turnaround time is not critical. For traditional methods such as t-bar tags, information can be had in close to real-time. TNT can only get “close” to that.

However, there are other advantages to TNT that substantially mitigate what little time might be lost there.

At the core of TNT is an extremely powerful computing system. TNT can receive radio-telemetered images, and process them automatically. We can process an image and determine if that image is of a previously "caught" animal in milliseconds.

Better/Cheaper/Faster enables more research within the constraints of existing budgets. Additionally, however, TNT facilitates eco-system-wide, and multi-species evaluations without imposing any constraints and only minor changes to current operating procedures by a diverse and uncoordinated group of scientists. Here’s an example:

As all of us who have been involved with natural resource management know (sometimes painfully so), getting large-scale studies to actually produce something useful is difficult. Imagine an eco-system such as Prince William Sound (PWS), Alaska. Biologists are currently "running hither and yon" collecting a broad spectrum of population data. Forcing them to share their data imposes a non-trivial layer of bureaucracy and labor on them. It is no wonder that many resist these coordination efforts. Often, a researcher’s government funding is tied to such cooperation. TNT eliminates any and all forced cooperation. In fact, there is no overt cooperation required at all, forced or otherwise. Each researcher simply continues to do what they have always done. The information comes into the TNT system, and is stored. After some interval of time, say three years, imagine the volume of data on PWS contained in the system. The geographic area covered is the whole Sound. The species covered run the list from benthic animals such as mollusks, crabs, and groundfish, to the pelagic fishes, to the surface species such as otters and sea lions, and to the air with birds and insects. At some point, someone can start mining this data and begin to obtain powerful answers to some broad-scoped questions. Most importantly, interested laymen become an army of field data collectors from which all researchers, from all "sides" of an issue, will benefit. All without a single researcher being burdened by extra bureaucracy, or being forced to coordinate their research with other researchers.

BIOPAR and the Biopar logo are registered trademarks of, and The Natural Tag and TNT are trademarks of Biopar, LLC. Patent No. 8,113,151.