(FORTUNE Magazine) – When he ran Canon's North American division, Fujio Mitarai enjoyed playing golf with Jack Welch at the Fairfield Country Club in Connecticut. He liked talking business with the then-CEO of General Electric too. "I respected his sense of strategy, his enthusiasm and sense of mission," says Mitarai, who spent 23 years in the U.S. But after Mitarai returned to Japan in 1989 and became president of Canon in 1995, he didn't embrace the GE way. "The management style of Mr. Welch is something that cannot be done in Japan," Mitarai says. "His style at his huge conglomerate was to have frequent acquisitions and disinvestments. That would be difficult for us."

A typical Japanese boss in denial about the need for change? Far from it. Mitarai has forged a hybrid management style that embraces the hardheaded American pursuit of profit with traditional Japanese business values. Neutron Jack would be stunned to hear Mitarai pledge a no-layoff policy. "Our company offers lifetime employment" within Japan, Canon's CEO says flatly. But Mitarai has not been shy about smashing together hitherto independent divisions, exiting money-losing product lines such as personal computers and liquid crystal displays, or expanding the scope of merit-based pay. The result: a Japanese company that is intensely competitive despite all the ink spilled over Japan's "lost decade." And now that competitive caliber is being aimed full-bore at the U.S. market--and especially at a troubled icon, Xerox.

Mention Canon and most people think cameras. The company remains a world-class maker of both digital and film cameras and camcorders. But the main engine of Canon's business is copiers, printers, and other office equipment, which together generate 77% of the company's revenues, compared with 13% from cameras. (The rest comes from semiconductor manufacturing equipment, medical systems, and other products.) Sales in the U.S. account for about a third of Canon's revenues and are expanding by some 7% a year.

Canon is the world's No. 1 copier company, with a 30% market share. But until recently it had focused on smaller and medium-sized machines. Now it is pushing into enterprise-wide document and image management--copiers, printers, faxes, scanners, all networked together--in other words, Xerox territory. Much as Japanese automakers are targeting Detroit's last bastions of profitability in pickup trucks and sport-utility vehicles, Canon is cracking the high-end market that Xerox had long dominated. Its share of the overall U.S. laser copier market grew from 13.3% in 1998 to 36.6% in unit sales last year. Meanwhile, Xerox's share slipped from 38.3% to 14.2%, according to International Data Corp., a Framingham, Mass., research firm. Canon has also moved ahead of Xerox in the high-speed digital copier market (although Xerox remains king of the fastest single-function printers). "Canon in the past year or two is winning," says Angele Boyd, an IDC analyst.

And it's winning on more than one front. Lexmark won't be able to find much relief in the inkjet printer market, where Canon is turning up the heat. Eastman Kodak is going to face a tougher challenge as Canon comes on strong in digital cameras and introduces peripherals that can print digital photos directly from cameras. Even GE's medical division could find stiff new competition from Canon in the management and storage of digital medical images such as X-rays. "At Canon they know exactly what their strengths are, and they keep concentrating on them," says Shin Horie, a Goldman Sachs analyst in Tokyo.

Canon's strength is all the more remarkable because so many other icons of Japanese high tech are floundering. Hitachi, Matsushita, Toshiba, NEC, and Fujitsu all posted losses in the fiscal year that ended in March. Canon earned $1.4 billion on revenues of $23.9 billion in 2001. And that was at a time when all of the world's economies were soft.

Beyond financials, the fundamentals of Canon's businesses are healthy as well. The company is second in the world to IBM in patents held. It has increased research and development spending from 6% of sales to 8% since Mitarai took over. And it is maintaining its production base in Japan rather than allowing itself to be hollowed out by shifting too much manufacturing to China or Southeast Asia. To accomplish all that, Mitarai, still vigorous at 67, is relentlessly driving down costs at his Japanese plants, in part by absorbing lean manufacturing techniques from the Japanese auto industry.

One sign of Canon's clout is that it is looking for an acquisition in the U.S. Mitarai is mum about the target, but the signs point to a software or services company that will enable Canon to manage documents and images across the Internet and other platforms, a goal everyone in the industry is chasing. The timing is ideal: Many U.S. tech firms have depressed stock prices, while Canon is sitting on about $3.8 billion in cash.

It's tricky for Mitarai to talk about what he assimilated in the U.S. because American management techniques aren't always admired in Japan. But it's clear he learned how to make decisions quickly and project his personality throughout a company, two traits most Japanese CEOs lack. Another big difference: Mitarai is focused on profit, not market share or size for the sake of size. "I have been influenced considerably by my experience in the United States," Mitarai says, sitting in his office in Canon's new, high-tech headquarters in Tokyo. (Nothing so primitive as smart cards to access floors--sensors read your fingerprints.) But he resists calling his management style American. "It's a style you would see globally," he argues.

However it's described, American CEOs with whom he interacts believe he is a different kind of Japanese business leader. "What you find in many Japanese companies is that the CEOs are often ceremonial, just figureheads," says Steve Appleton, head of semiconductor maker Micron Technology, a Canon customer and supplier. "That's not the case with Mitarai at all. He's active in leading the company. He's pretty driven."

Mitarai's style is shaped by his keen interest in technology. The company was founded by a medical doctor--his uncle Takeshi Mitarai--in the 1930s with the goal of responding to the technological marvels of Germany's Leica camera company. When Canon decided to enter the copier business in 1962, it encountered a patent wall that had been erected by then-dominant Xerox around its xerography system. That was a watershed moment for Canon. It took the company three years to create its own copying technology and another three to develop its first product, but the effort launched it on the path to world-class R&D.

Mitarai has taken that tradition and pushed it further. Canon boasts a total of 74,000 U.S. patents--2,000 just for the photosensitive drum in a digital copier. Unlike patent leader IBM, which has scientists pondering the nature of outer space, Canon has engineers designing and redesigning products, so its patents are concentrated in fields more tightly connected to its businesses.

Boosting spending on research and filing patents is fine, but is Mitarai getting enough bang for his yen? Satomi Ushioda, an analyst at Nikko Salomon Smith Barney in Tokyo, is skeptical. She notes that the company hasn't introduced a product based on entirely new technology since it invented the inkjet printer in the 1990s. She thinks Canon needs a breakthrough product to fuel growth. "Even though they spend all that money every year, they cannot find a new earnings driver," says Ushioda. "They can't use the patents to make money."

Mitarai says one breakthrough may be on the way. Canon's new display technology, surface-conduction electron-emitter display, or SED, could be a $1 billion-plus product, he says. It would consume less power than today's displays and allow thinner, bigger, wall-mounted products. Mitarai launched a joint venture with Toshiba three years ago to see if the technology could be mass-produced, and a prototype is now in hand. He hopes to know within a year whether it can be manufactured profitably. If so, it would allow Canon not only to capture images through cameras, scanners, and other input devices but also to display them on screens that it co-manufactures.

Even if that bet doesn't pay off, Mitarai says the way Canon pushes incremental improvements into products is at the heart of the company's success. More than 60% of the products it sells today have been introduced within the past two years. Xerox officials quarrel with that figure, arguing that some of the products Canon counts as new are based on existing platforms. Still, it's clear that Canon's R&D is as good as that of any of its rivals. That's how it can make money in mature sectors with thin margins. "We can avoid having to reduce prices by launching products one after another," says Mitarai, "products that are competitive and also contain higher and higher value-added." Canon also has been a leader in creating multifunction machines that combine scanning, printing, copying, and other functions. "We are now in an age," he says, "when the technology will change even if we are producing the same kind of products."

Despite its successes, Canon knows it must do better. Katsuichi Shimizu, deputy chief executive in charge of office imaging products for Canon, acknowledges that it's more difficult to shift Japanese scientists from one project to another than it is in the U.S. And it's harder to inject outsiders into Canon's R&D shop, just as it is at most Japanese companies. "That's one reason Japanese companies are behind the Americans," Shimizu says.

To narrow the gap, Canon conducts research at two facilities in California, where engineers, most of them American, work on the software and systems-integration issues critical for the company to sell more systems to big corporate users. The idea is to help Canon move from being a manufacturer of "boxes" to a company that understands complex corporate environments with large computer systems and communications networks, including the Internet. "To make sales in these large enterprises, what Canon offers has to work well with what's already installed," says Toru Takahashi, chief technical officer for Canon USA, who is in charge of the California facilities.

Of course, R&D means nothing if it can't be turned into products. Canon executives are openly scornful of how Xerox ignored ideas coming out of its own Palo Alto Research Center a couple of decades ago. One problem was the geographic and mental distance between researchers and the top brass. Mitarai avoids that problem in part by maintaining a large production base in Japan, even though costs are lower overseas and 70% of his sales are outside Japan. He says the two-way flow of ideas between his researchers and production staff is essential to maintaining development speed. American companies, he says, outsource too much of their production.

To get a glimpse of how Mitarai is defending his Japanese factories, travel to the Ami plant in Ibaraki prefecture, a largely agricultural area two hours north of Tokyo. The workforce is predominantly women in their 30s, and there is almost no turnover, all of which makes it a reasonably low-cost production base in a high-cost country. Still, the price of labor is as much as 20 times higher than in China, the plant's managers estimate.

So the company has to get more productivity out of its Japanese workforce. At the Ami plant, as at its other factories, Canon has ripped out old conveyor belts and heavy equipment that was once bolted to the floor. Starting in 1998, it replaced traditional assembly lines with what Canon calls "cell" production. Small teams, or cells, of typically six workers concentrate on building a single copying machine. Production equipment is lighter and more modular than before, and can be more easily reconfigured.

Workers also have been encouraged to come up with their own solutions. They design their own seats, for example, so that as they squat like baseball catchers working on a copier, all their work is in the strike zone. One worker rigged a lid that comes down over photosensitive drums she is installing into copiers to prevent dust and light from harming them. The result is a precisely orchestrated whirl of hands and arms. "It's a small-brain approach, not a big-brain approach," says Goldman Sachs's Horie.

These steps have dramatically improved productivity. The Ami facility cut 1,200 workers (deployed elsewhere, in line with Canon's no-layoff policy) and eliminated seven parts warehouses, reducing costs by half and boosting productivity by more than 20%. That kind of improvement has helped Mitarai maintain 65% of Canon's production in Japan. He may shift another 5% abroad by 2005, but he's not in a hurry to do that. "Whatever the trend in exchange rates or whatever the external factors," Mitarai says, "a manufacturing company is always faced with the mission of transforming itself into a company that can produce higher value-added to absorb the increase in the cost of living in the country it's operating in."

Non-Japanese managers often say that lifetime employment is inflexible and fosters complacency, but that doesn't seem to be the case at Canon. Mitarai argues that the system of incentives he has pushed means Canon rewards on the basis of performance, not just seniority. The system isn't based on operating results but rather performance on a series of tests beginning at age 25. If you do well on the tests, you get promotions and make more money than your peers--as much as 80% more. This isn't radical by Western standards, but it does make Canon different from most Japanese companies.

That Mitarai can keep coming up with more-sophisticated products while driving costs down means trouble for Xerox. Canon has already run Xerox out of the low end of the copier market. (Xerox says it intentionally retreated from this segment of the market to focus on maintaining its hold on the top rungs.) Now it is applying the same formula to the high-end corporate document-handling business. For example, analysts divide the copier market into six segments, depending on speed, and in the top two tiers for digital copiers--more than 70 pages a minute--Canon's market share in the U.S., based on dollar sales, increased from nothing in 2000 to 23% last year, according to IDC.

Xerox's share in those two segments increased from 12% to 17%, but it has good reason to be worried. Last year Xerox and Canon went head to head for a contract with Coca-Cola Bottling of North Carolina. The bottler's goal was to rip out an expensive, inefficient gaggle of copiers, printers, and fax machines made by HP, Canon, Brother, and Xerox, and replace them with one multifunction product that could do it all while communicating with the company's mainframe through a wide-area network. It was the type of deal that is Xerox's bread and butter.

Xerox fought for the business, sending in a senior sales executive to assure Coca-Cola Bottling that its machines could handle the job. "But I never saw him again," complains Mike Perkis, the bottler's vice president of materials management. That made him worry that Xerox's management upheavals and accounting problems would prevent it from providing enough customer support. So his company decided to lease Canon's multifunction ImageRunners for more than 100 locations--a contract worth in excess of $2 million a year, but one that Coca-Cola Bottling says saved it up to 10% in equipment costs.

"There are wins and losses every day," says Christa Carone, a Xerox spokeswoman, adding that Xerox remains the main vendor for Coca-Cola Co. "Our business has strengthened considerably over the past 18 months." Among other things, Xerox is launching new products to defend its turf. It has introduced DocuTech, a new high-speed, high-volume publishing system that could expand the company's share in printing shops. It is also marketing iGen3, a high-end digital production system for color printing. Gilbert Hatch, president of Xerox Office Systems Group in Rochester, N.Y., says Xerox has had management stability since Anne Mulcahy became CEO two years ago. "Our office strategy is moving in the right direction," Hatch says. "We are focused like a laser beam on our competition."

Perhaps so. But once it was Xerox in the driver's seat, with upstart Canon struggling to prove its credibility. Now it's the other way around. Unless Xerox can pull off a textbook-perfect revitalization, the Canon that Mitarai has built will keep eating into the American company's high-profit businesses. Seen through the eyes of Mitarai and other Canon executives, the next stage of their 40-year war against Xerox is tilting in their favor.